Monolithic integrated voltage divider device based on depletion field effect
By using the design of lateral bipolar junction transistors and metal oxide semiconductor transistors in monolithic integrated circuits, combined with depletion field effect voltage dividers, the manufacturing problems of bipolar, CMOS and DMOS devices in monolithic integration are solved. , achieving high performance, low cost and high efficiency voltage reduction.
Patent Information
- Application Number
- CN202380034580.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2022-02-25
- Filing Date
- 2023-02-23
- Publication Date
- 2024-11-26
AI Technical Summary
The prior art is difficult to efficiently collaborate in monolithic integrated circuits in efficient integrated circuits, resulting in complex manufacturing, high cost, low reliability, and difficult to reduce the effective voltage between low voltage and high voltage regions.
Designed with transverse bipolar junction transistors (L-BJT) and metal oxide semiconductor transistors, physical isolation of low-voltage and high-voltage areas is achieved by sharing manufacturing steps and structural features, and using depletion field effect voltage dividers ( DFE-PD) Scales the high voltage signal to the low voltage signal.
It realizes high-performance, low-cost monolithic integrated circuits, reduces manufacturing complexity and cycle time, improves device reliability and voltage reduction efficiency, and adapts to a wide range of application voltage nodes.
Smart Images

Figure CN119032424A_ABST
Abstract
Description
[0001] Incorporation by reference into any priority application
[0002] Any and all applications for which foreign or domestic priority claims are identified in the Application Data Sheet filed with the present application are incorporated herein by reference pursuant to 37 CFR 1.57.
[0003] This application claims priority to U.S. Provisional Application No. 63 / 268,585, filed on February 25, 2022, U.S. Provisional Application No. 63 / 268,590, filed on February 25, 2022, U.S. Provisional Application No. 63 / 268,588, filed on February 25, 2022, U.S. Provisional Application No. 63 / 268,587, filed on February 25, 2022, and U.S. Provisional Application No. 63 / 268,519, filed on February 25, 2022. The entire contents of each application cited in this paragraph are hereby incorporated herein by reference for all purposes and made a part of this specification. Technical Field
[0004] The disclosed technology relates generally to integrated circuits and, more particularly, to monolithically integrated high performance transistors and high voltage devices. Background Art
[0005] Some integrated circuit (IC) devices include a combination of one or more different types of semiconductor devices, such as complementary metal oxide semiconductor (CMOS) devices, bipolar devices, and double diffused metal oxide semiconductors. Bipolar CMOSDMOS (BCDMOS) devices refer to IC devices that include features of each of bipolar, CMOS, and DMOS devices and can combine their features. For example, bipolar devices can provide precise analog functions, CMOS devices can provide digital designs, and DMOS devices can provide power and high-voltage components. The combination of these different device technologies can bring about a combination of these and other advantages, including high-speed and high-voltage operation, improved reliability, reduced electromagnetic interference, and smaller chip area. BCDMOS devices have been used in a wide range of products and applications in areas such as power management, analog data acquisition, and power actuators. Summary of the invention
[0006] In one aspect, an integrated circuit (IC) device includes: a metal oxide semiconductor (MOS) transistor and a bipolar junction transistor (BJT), the MOS transistor including a gate stack formed on its channel region, and the BJT including a layer stack formed on its collector region. The gate stack and the layer stack have one or more corresponding layers with common physical dimensions.
[0007] In another aspect, an integrated circuit (IC) device includes: a metal oxide semiconductor (MOS) transistor and a bipolar junction transistor (BJT), the MOS transistor including a gate stack formed on its channel region, and the BJT including a layer stack formed on its collector region. The gate stack and the layer stack have corresponding spacer layer structures formed on their sidewalls having common physical dimensions.
[0008] In another aspect, an integrated circuit (IC) device includes: a metal oxide semiconductor (MOS) transistor and a bipolar junction transistor (BJT), the MOS transistor including a gate stack formed on its channel region, and the BJT including a layer stack formed on its collector region. The MOS transistor and the BJT have corresponding implant diffusion regions having a common implant dopant distribution.
[0009] In another aspect, an integrated circuit (IC) device includes a metal oxide semiconductor (MOS) transistor and a bipolar junction transistor (BJT) formed in a common semiconductor substrate. The BJT includes a collector region electrically connected in series to a high voltage divider substrate region and physically isolated therefrom by an isolation structure. The BJT also includes a layer stack including a conductive field plate formed on the collector region.
[0010] In another aspect, an integrated circuit (IC) device includes a metal oxide semiconductor (MOS) transistor and a bipolar junction transistor (BJT) formed in a common semiconductor substrate. The BJT includes a collector region electrically connected in series to a high voltage divider substrate region and physically isolated therefrom by an isolation structure. The collector region and the high voltage divider substrate region are electrically connected by one or more metallization layers formed above a main surface of the common semiconductor substrate.
[0011] In another aspect, an integrated circuit (IC) device includes a metal oxide semiconductor (MOS) transistor and a bipolar junction transistor (BJT) formed in a common semiconductor substrate. The BJT includes a collector region electrically connected in series to a high voltage divider substrate region and physically isolated therefrom by an isolation structure. The high voltage divider substrate region is configured to reduce more than 50% of a combined voltage applied across the BJT and the high voltage divider substrate region.
[0012] In another aspect, an integrated circuit (IC) device includes a metal oxide semiconductor (MOS) transistor and a lateral double diffused metal oxide semiconductor (LDMOS) transistor formed in a common semiconductor substrate. The LDMOS transistor includes an extended drain drift region physically separated from a channel region of the LDMOS transistor by an isolation structure. The gate stack of the MOS transistor and the gate stack of the LDMOS transistor have one or more corresponding layers having common physical dimensions.
[0013] In another aspect, an integrated circuit (IC) device includes a metal oxide semiconductor (MOS) transistor and a lateral double diffused metal oxide semiconductor (LDMOS) transistor formed in a common semiconductor substrate. The LDMOS transistor includes an extended drain drift region physically separated from a channel region of the LDMOS transistor by an isolation structure. The MOS transistor and the LDMOS transistor have an implanted diffusion region having a common implant dopant distribution.
[0014] In another aspect, an integrated circuit (IC) device includes a metal oxide semiconductor (MOS) transistor and a lateral double diffused metal oxide semiconductor (LDMOS) transistor formed in a common semiconductor substrate. The LDMOS transistor includes an extended drain drift region physically separated from a channel region of the LDMOS transistor by an isolation structure. The extended drain drift region includes a conductive field plate formed thereon.
[0015] In another aspect, an integrated circuit (IC) device includes a metal oxide semiconductor (MOS) transistor and one or both of a bipolar junction transistor (BJT) and a lateral double diffused metal oxide semiconductor (LDMOS) transistor formed in a common silicon substrate. The IC device also includes a high voltage divider region formed above a major surface of the common semiconductor substrate and integrated in a back end (BEOL) of the IC device. The high voltage divider region is formed of a wide bandgap semiconductor material.
[0016] In another aspect, an integrated circuit (IC) device includes a metal oxide semiconductor (MOS) transistor and one or both of a bipolar junction transistor (BJT) and a lateral double diffused metal oxide semiconductor (LDMOS) transistor formed in a common silicon substrate. The IC device also includes a high voltage divider region formed above a major surface of the common semiconductor substrate and integrated in a back end (BEOL) of the IC device. The high voltage divider region includes a main portion that includes a lightly doped semiconductor region.
[0017] In another aspect, an integrated circuit (IC) device includes a metal oxide semiconductor (MOS) transistor and one or both of a bipolar junction transistor (BJT) and a lateral double diffused metal oxide semiconductor (LDMOS) transistor formed in a common silicon substrate. The IC device also includes a high voltage divider region formed above a major surface of the common semiconductor substrate and integrated in the back end (BEOL) of the IC device. In operation, the high voltage divider region is configured to drop a higher voltage at least relative to the MOS transistor.
[0018] In another aspect, a lateral bipolar junction transistor (BJT) includes an emitter region formed in a base well, a lightly doped collector region separated from the base well by a drift region in a lateral direction, the drift region being doped with the same type of dopant at a lower dopant concentration relative to the lightly doped collector region. The lateral BJT also includes a layer stack formed on the drift region, wherein a boundary of the base well or the emitter region is aligned with an edge of the layer stack in a vertical direction intersecting the lateral direction.
[0019] In another aspect, a lateral bipolar junction transistor (BJT) includes an emitter region formed in a base well, a lightly doped collector region separated from the base well in a lateral direction by a drift region, the drift region being doped with the same type of dopant at a lower dopant concentration relative to the lightly doped collector region. The lateral BJT also includes a layer stack formed on the drift region, with a spacer layer formed on a sidewall thereof, wherein a base length of the lateral BJT extends in a lateral direction between the emitter region and the drift region, and is defined by a lateral width of a bottom of the spacer layer formed on the sidewall of the layer stack.
[0020] In another aspect, a lateral bipolar junction transistor (BJT) includes an emitter region formed in a base well, and a heavily doped collector region spaced apart from the base well in a lateral direction, the heavily doped base region being spaced apart by a drift region doped with the same type of dopant at a lower dopant concentration relative to the heavily doped collector region. The lateral BJT also includes a layer stack formed on at least a first lateral portion of the drift region, the layer stack including a conductive reduced surface field (RESURF) layer.
[0021] In another aspect, an integrated circuit (IC) device includes a metal oxide semiconductor (MOS) transistor including a heavily doped (HD) drain region, a gate stack, a drain region extension extending laterally from the HD drain region to an edge of the gate stack, and at least one drain field plate extending over at least 20% of the length of the drain region extension. In addition, the integrated circuit (IC) device includes a bipolar junction transistor (BJT) including a heavily doped (HD) collector region, a layer stack, a drift region extension extending laterally from the HD collector region to an edge of the layer stack, and at least one field plate extending over at least 20% of the drift region extension, wherein the at least one drain field plate and the at least one field plate have at least one common physical dimension.
[0022] In another aspect, an integrated circuit (IC) device includes a metal oxide semiconductor (MOS) transistor including a heavily doped (HD) drain region, a gate stack, a drain region extension extending laterally from the HD drain region to an edge of the gate stack, and a thick dielectric layer extending on the drain region extension. In addition, the integrated circuit (IC) device includes a bipolar junction transistor (BJT) including a heavily doped (HD) collector region, a layer stack, a drift region extension extending laterally from the HD collector region to an edge of the layer stack, and a thick isolation dielectric layer extending on the drift region extension, wherein the thick dielectric layer and the thick isolation dielectric layer have at least one common physical dimension.
[0023] In another aspect, an integrated circuit (IC) device includes a metal oxide semiconductor (MOS) transistor including a heavily doped (HD) drain region, a gate stack, and a drain region extension extending laterally from the HD drain region to an edge of the gate stack, the gate stack including a gate layer extending over a gate dielectric layer and a portion of the drain region extension. In addition, the integrated circuit (IC) device includes a bipolar junction transistor (BJT) including a heavily doped (HD) collector region, a layer stack, and a drift region extension extending laterally from the HD collector region to an edge of the layer stack, and a reduced surface field (RESURF) layer extending over a RESURF dielectric layer and a portion of the drift region extension, wherein the gate layer and the RESURF layer have at least one common physical dimension.
[0024] In another aspect, a bipolar junction transistor (BJT) is formed in a substrate, the BJT including a base well, an emitter region formed in the base well, a heavily doped (HD) collector region separated from the base well in a lateral direction by a drift region, and a first conductive field plate disposed on a first lateral portion of the drift region, the drift region laterally spanning more than 20% of a length of the drift region, wherein the emitter region, the drift region, and the HD collector region have the same polarity opposite to the polarity of the base well.
[0025] In another aspect, a bipolar junction transistor (BJT) is formed in a substrate, the BJT including a base well, an emitter region formed in the base well, and a heavily doped (HD) collector region laterally separated from the base well by a drift region in which an HD collector region is formed. The drift region has a lower dopant concentration relative to the HD collector region and the base well, and the emitter region, the drift region, and the HD collector region have the same polarity opposite to the base well polarity.
[0026] In another aspect, a bipolar junction transistor (BJT) is formed in a substrate, the BJT including a base well, an emitter formed in the base well, a heavily doped (HD) collector region spaced apart from the base well in a lateral direction by a drift region, and a thick dielectric layer disposed on a first lateral portion of the drift region, the first lateral portion being laterally longer than 20% of a length of the drift region, the emitter region, the drift region, and the HD collector region having the same polarity opposite to the polarity of the base well.
[0027] In another aspect, an integrated circuit (IC) device configured to reduce a voltage between an input and an output, wherein the IC device includes an isolation substrate region formed in a semiconductor substrate while being electrically isolated from the semiconductor substrate in vertical and lateral directions, wherein a plurality of alternating doped regions are formed in the isolation substrate region, wherein the alternating doped regions are laterally arranged in a lateral direction and are alternately doped with opposite types of dopants. The alternating doped regions include an input drift region and an output drift region, each doped with a first type of dopant, wherein the input drift region is connected to an input and the output drift region is connected to an output, an inter-gate region and a substrate region of the isolation substrate region, each doped with a second type of dopant, wherein the inter-gate region is laterally interposed between the input and output drift regions. A heavily doped first gate region is formed in the inter-gate region, wherein the input drift region is elongated to have a first lateral length along the lateral direction, the first lateral length being at least twice longer than a second lateral length of the inter-gate region.
[0028] In another aspect, an integrated circuit (IC) device configured to reduce a voltage between an input and an output, wherein the IC device includes an isolation substrate region formed in a semiconductor substrate while being electrically isolated from the semiconductor substrate in vertical and lateral directions, wherein a plurality of alternating doped regions are formed in the isolation substrate region, wherein the alternating doped regions are laterally arranged in a lateral direction and are alternately doped with opposite types of dopants. The alternating doped regions include an input drift region and an output drift region, each doped with a first type of dopant, wherein the input drift region is connected to an input and the output drift region is connected to an output, and an inter-gate region and a substrate region of the isolation substrate region, each doped with a second type of dopant, wherein the inter-gate region is laterally interposed between the input and output drift regions. The inter-gate region has a dopant concentration lower than the doping concentrations of the input drift region, the output drift region, and the substrate region, and a heavily doped first gate region formed in the inter-gate region.
[0029] In another aspect, an integrated circuit (IC) device configured to reduce a voltage between an input and an output, wherein the IC device includes an isolation substrate region formed in a semiconductor substrate and electrically isolated from the semiconductor substrate in vertical and lateral directions, wherein a plurality of alternating doped regions are formed in the isolation substrate region, wherein the alternating doped regions are laterally arranged in a lateral direction and are alternately doped with opposite types of dopants. The alternating doped regions include an input drift region and an output drift region, each doped with a first type of dopant, wherein the input drift region is connected to an input and the output drift region is connected to an output, and an inter-gate region and a substrate region of the isolation substrate region, each doped with a second type of dopant. The inter-gate region is laterally between the input and output drift regions. In addition, the integrated circuit (IC) device includes a dielectric layer covering the input drift region and at least one conductive field plate extending laterally above the dielectric layer, and a heavily doped (HD) first gate region formed in the inter-gate region.
[0030] In another aspect, an integrated circuit (IC) device includes: at least one low voltage active semiconductor device formed in a first substrate region of a semiconductor substrate; and a depletion field effect voltage divider (DFE-PD) formed in a second substrate region of the semiconductor substrate and configured to reduce a high voltage input signal received from a high voltage node to generate a low voltage output signal, and provide the low voltage output signal as an input signal to the low voltage active semiconductor device. The low voltage active semiconductor device and the DFE-PD are physically separated while being electrically connected in series, and the first and second substrate regions are electrically isolated by an isolation structure.
[0031] In another aspect, an integrated circuit (IC) device includes: at least one low voltage active semiconductor device formed in a first substrate region of a semiconductor substrate; and a depletion field effect voltage divider (DFE-PD) formed in a second substrate region of the semiconductor substrate and configured to reduce a high voltage input signal received from a high voltage node to generate a low voltage output signal, and provide the low voltage output signal as an input signal to the low voltage active semiconductor device. The DFE-PD includes a plurality of alternately doped regions that are laterally arranged in a lateral direction and alternately doped with opposite types of dopants. The alternately doped regions include an input drift region and an output drift region, each doped with a first type of dopant, wherein the input drift region is connected to an input, and the output drift region is connected to an output, and a gate inter-region laterally between the input and output drift regions, wherein the gate inter-region and the second substrate region are doped with a second type of dopant, and a heavily doped first gate region formed in the gate inter-region.
[0032] In another aspect, an integrated circuit (IC) device includes: at least one low voltage active device formed in a semiconductor substrate; and a depletion field effect voltage divider (DFE-PD) formed above a major surface of the semiconductor substrate and separated therefrom by a dielectric layer, the DFE-PD being configured to scale a high voltage input signal received from a high voltage node into a low voltage output signal and provide the low voltage signal as an input signal to the low voltage active device. The DFE-PD includes a doped semiconductor region and a depletion control region, the doped semiconductor region extending in a lateral direction parallel to the major surface of the substrate between a heavily doped (HD) input region and a heavily doped (HD) output region, the depletion control region being formed in the doped semiconductor region and being counter doped relative to the remainder of the doped semiconductor region, wherein the DFE-PD and the low voltage active device are electrically connected by an electrical connection.
[0033] In another aspect, an integrated circuit (IC) device includes at least one low voltage active device formed in a semiconductor substrate; and a depletion field effect voltage divider (DFE-PD) formed above a major surface of the semiconductor substrate and isolated therefrom by a dielectric layer, the DFE-PD being configured to scale a high voltage input signal received from a high voltage node into a low voltage output signal and provide the low voltage signal as an input signal to the low voltage active device. The DFE-PD includes a doped semiconductor region extending laterally along a lateral direction parallel to the major surface of the substrate between a heavily doped (HD) input region receiving the low voltage output signal and a heavily doped (HD) output region outputting the low voltage signal. In addition, the DFE-PD includes a depletion control region formed in the doped semiconductor region and counter doped relative to the rest of the doped semiconductor region, and at least one conductive field plate extending laterally on the doped semiconductor region and electrically connected to the low voltage active device through an electrical connection formed at least partially above the semiconductor substrate. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] The above overview and the following detailed description of the exemplary embodiments can be better understood when read in conjunction with the accompanying drawings. In order to illustrate the present disclosure, exemplary configurations of the present disclosure are shown in the accompanying drawings. In addition, it will be understood by those skilled in the art that the drawings are not drawn to scale. Where possible, the same elements are represented by the same numerals. The detailed description of the embodiments and the embodiments set forth in the accompanying drawings present various descriptions of specific embodiments of the present invention. However, the present invention can be embodied in a variety of different ways. It should be understood that some embodiments may include more elements than shown in the drawings and / or a subset of the elements shown in the drawings. In addition, some embodiments may combine any suitable combination of features from two or more drawings. The present disclosure is not limited to the specific methods and devices disclosed herein.
[0035] Embodiments of the present disclosure will now be described, by way of example only, with reference to the following patent drawings, in which:
[0036] Figure 1A A multi-die electronic platform is schematically shown including bipolar, CMOS, and DMOS devices mounted as multiple dies on a single substrate.
[0037] Figure 1B is a block diagram illustrating the monolithically integrated BCDMOS (or superset) design concept.
[0038] Figure 2 Schematically illustrated is a side cross-sectional view of a bipolar CMOS (BiCMOS) integrated circuit (IC) according to certain embodiments disclosed herein.
[0039] Figure 3A A side cross-sectional view of a MOS transistor with an extended drain drift region and a field plate is schematically shown.
[0040] Figure 3B Schematically illustrated is a side cross-sectional view of a lateral bipolar junction transistor (BJT) with an extended drift region and a field plate according to certain embodiments disclosed herein.
[0041] Figure 4A Schematically shown is a side cross-sectional view of a MOS transistor with an extended drain region and close-up views of the source region (left) and the extended drain region (right).
[0042] Figure 4B Schematically illustrates a side cross-sectional view of a lateral bipolar junction transistor (L-BJT) with an extended drift region and close-up views of the base / emitter region (left) and the extended collector region (right) according to certain embodiments disclosed herein.
[0043] Figure 5A Schematically illustrated is a side cross-sectional view of a lateral isolated voltage divider (PD) electrically connected to a MOSFET according to certain embodiments disclosed herein.
[0044] Figure 5B Schematically illustrated is a side cross-sectional view of a laterally isolated PD electrically connected to a vertical bipolar junction transistor (V-BJT) according to certain embodiments disclosed herein.
[0045] Figure 5C Schematically illustrated is a side cross-sectional view of a lateral isolated voltage divider electrically connected to a lateral bipolar junction transistor (L-BJT) according to certain embodiments disclosed herein.
[0046] Fig. 6A Schematically illustrated is a three-dimensional (3D) view of a vertically isolated PD according to certain embodiments disclosed herein.
[0047] Figure 6BA 3D view of a PD fabricated in the back end of line (BEOL) of an IC relative to a substrate including a BCDMOS device electrically connected to the PD is schematically shown.
[0048] Fig. 7A Schematically illustrated is a side cross-sectional view of an example L-BJT according to certain embodiments disclosed herein.
[0049] Figure 7B A side cross-sectional view of a MOS transistor is schematically shown.
[0050] Figure 7C A side cross-sectional view of another disclosed embodiment of an L-BJT is schematically shown.
[0051] Fig.7D A side cross-sectional view of a LDMOS transistor is schematically shown.
[0052] Fig. 8A A 3D view of another disclosed embodiment of an L-BJT is schematically shown.
[0053] Figure 8B Schematically shows Fig. 8A Top view of the L-BJT shown in .
[0054] Figure 8C A side cross-sectional view of another embodiment of an L-BJT is schematically shown.
[0055] Fig.8D Schematically shows Figure 8C Top view of the L-BJT shown in .
[0056] Fig. 9A Schematically illustrated is a side cross-sectional view of an L-BJT with a dedicated base spacer layer according to certain embodiments disclosed herein.
[0057] Fig. 9B Schematically illustrated is a side cross-sectional view of an L-BJT having an elongated emitter region according to certain embodiments disclosed herein.
[0058] Fig. 9C Schematically illustrated is a side cross-sectional view of an L-BJT having a vertical emitter portion according to certain embodiments disclosed herein.
[0059] Fig.9D Schematically illustrated is a side cross-sectional view of an L-BJT having an elongated emitter region and a vertical emitter region according to certain embodiments disclosed herein.
[0060] Fig.9E Schematically illustrated is a side cross-sectional view of a vertically and laterally isolated L-BJT according to certain embodiments disclosed herein.
[0061] Fig.9F Schematically illustrated is a side cross-sectional view of an L-BJT having a heat-diffused base region according to certain embodiments disclosed herein.
[0062] Figures 10A-10E Selected fabrication steps for fabricating an L-BJT according to certain embodiments disclosed herein are schematically illustrated.
[0063] Fig.10F Schematically illustrated are fabrication steps for fabricating an L-BJT having an elongated emitter region according to certain embodiments disclosed herein.
[0064] Figures 11A-11E Selected fabrication steps for fabricating an L-BJT according to certain other embodiments disclosed herein are schematically illustrated.
[0065] Fig.11F Schematically illustrated are fabrication steps for fabricating an L-BJT having an elongated emitter region according to certain other embodiments disclosed herein.
[0066] Figures 12A-12C Selected fabrication steps for fabricating an L-BJT having a vertical emitter portion according to certain embodiments disclosed herein are schematically illustrated.
[0067] Figures 12D-12F Selected fabrication steps for fabricating an L-BJT having a vertical emitter partial region according to certain other embodiments disclosed herein are schematically illustrated.
[0068] Figures 13A-13E Selected fabrication steps for fabricating an L-BJT having a thermally diffused base region according to certain embodiments disclosed herein are schematically illustrated.
[0069] Figures 14A-14E Selected fabrication steps for fabricating an L-BJT having a thermally diffused base region according to certain embodiments disclosed herein are schematically illustrated.
[0070] Fig.15A Schematically shows Fig.9D A cross-sectional view of an L-BJT is shown along with the voltage and electric field variations along its drift region.
[0071] Fig. 15B Schematically shows Figure 7C A cross-sectional view of an L-BJT is shown along with the voltage and electric field variations along its drift region.
[0072] Fig.16AA cross-sectional view of an embodiment of an L-BJT with an extended drift region having three field plates and a schematic dopant profile thereof is schematically shown.
[0073] Fig. 16B Schematically shows Fig.16A A cross-sectional view of an L-BJT is shown along with the electric field variation in the drift region along its extension.
[0074] Fig. 16C Schematically shown is a cross-sectional view of an L-BJT having an extended drift region with four field plates thereon, and the variation of voltage and electric field along the drift region extending therefrom.
[0075] Fig.17 Schematically illustrated are graphs showing the approximate operating voltages supported by several L-BJT designs as a function of the corresponding drift region lengths.
[0076] Fig.18 Schematically shown is a side cross-sectional view of an L-BJT having an extended drift region fully isolated using a buried oxide layer and trench isolation.
[0077] Fig.19 Schematically shown is a side cross-sectional view of an L-BJT with an isolated extended drift region fully isolated using trench isolation.
[0078] Fig. 20 The schematic diagram shows the input voltage at three different amplitudes. Fig.19 The depletion region of an L-BJT is shown.
[0079] Fig.21 An L-BJT with an extended drift region and a vertically separated drift region extension is schematically shown.
[0080] Fig.22A Schematically illustrated is a cross-sectional view of a semiconductor device structure configured to drop a voltage in the event of output saturation, and a corresponding voltage transfer function in an open-circuit output configuration.
[0081] Fig. 22B A cross-sectional view of another semiconductor device structure configured for voltage reduction without output saturation and with a slope change and a corresponding voltage transfer function in an open output configuration is schematically shown.
[0082] Fig. 22C A cross-sectional view of another semiconductor device structure configured for linear or near-linear voltage drop with a non-zero punch-through voltage and a corresponding voltage transfer function in an open-end output configuration is schematically shown.
[0083] Fig.22D A cross-sectional view of another semiconductor device structure configured for linear or near-linear voltage drop at zero punch-through voltage and a corresponding voltage transfer function in an open-end output configuration is schematically shown.
[0084] Fig. 22C A cross-sectional view of another semiconductor device structure for voltage reduction with a non-zero punch-through voltage and a corresponding voltage transfer function in an open-circuit output configuration is schematically shown.
[0085] Figures 23A-23B The voltage transformation through the inter-gate region and the distribution and evolution of the charge and electric field in the inter-gate region are schematically shown.
[0086] Figures 24A-24B Long (A) and short (B) inter-gate regions are schematically illustrated, showing the effect of the length of the inter-gate region on the corresponding voltage transition slope.
[0087] Fig.25A A cross-sectional view of a depletion field effect voltage divider (DFE-PD) with an extended drift region is schematically shown.
[0088] Fig.25B is for various values of input voltage, along Fig.25A Graph showing the voltage variation of the DFE-PD.
[0089] Fig.26A is a graph of the output voltage generated by the DFE-PD as a function of the input voltage provided to the DFE-PD.
[0090] Fig.26B is a graph of the output current per unit length of the DFE-PD as a function of the input voltage provided to the DFE-PD.
[0091] Fig.27A A cross-sectional view of a DFE-PD with a gate region on the input side is schematically shown.
[0092] Fig.27B A cross-sectional view of a DFE-PD with gate regions on the input and output sides is schematically shown.
[0093] Fig.28A Schematically shown is a side cross-sectional view of a DFE-PD with a field plate on the input region fabricated on a substrate and the voltage varying along the DFE-PD for several values of the input voltage.
[0094] Fig.28B is a graph of the output voltage generated by the DFE-PD as a function of the input voltage provided to the DFE-PD.
[0095] Fig.28C is a graph of the output current per unit length of the DFE-PD as a function of the input voltage provided to the DFE-PD.
[0096] Fig.28D Schematically shows Fig.28A The drift region, depletion control region, and depletion region in the output region of the DFE-PD are shown as a function of input voltage.
[0097] Fig.29 A side cross-sectional view of another DFE-PD with field plates on the input and output regions is schematically shown.
[0098] Fig. 30A A cross-sectional view of a laterally isolated DFE-PD electrically connected to a lateral bipolar junction transistor (L-BJT) according to certain embodiments disclosed herein is schematically shown.
[0099] Fig. 30B A cross-sectional view of a laterally isolated DFE-PD electrically connected to a MOS transistor is schematically illustrated according to certain embodiments disclosed herein.
[0100] Fig.31 A cross-sectional view of an example high voltage device is schematically shown, the device including electrical connections between a DFE-PD and an L-BJT fabricated on a common substrate, wherein the DFE-PD is electrically connected to the L-BJT while being laterally isolated from each other and vertically isolated from the substrate by a buried dielectric layer.
[0101] Fig.32 A cross-sectional view of an example high voltage device is schematically shown, the device including electrical connections between a DFE-PD and an L-BJT fabricated on a common substrate, wherein the DFE-PD is electrically connected to the L-BJT while being laterally isolated from each other and vertically isolated and confined by a buried dielectric layer.
[0102] Fig.33 A cross-sectional view of an example high voltage device including a DFE-PD vertically separated from a substrate and directly connected to an L-BJT is schematically shown.
[0103] Fig.34 A cross-sectional view of an example high voltage device including a DFE-PD vertically separated from a substrate and electrically connected to an L-BJT through conductive lines is schematically shown.
[0104] Fig.35 A cross-sectional view of an example high voltage device including a DFE-PD that is laterally and vertically isolated from a substrate and electrically connected to an L-BJT through conductive lines is schematically shown.
[0105] Fig.36 A cross-sectional view of another example high voltage device is schematically shown, the device including a DFE-PD vertically isolated from a substrate and electrically connected to an L-BJT through conductive lines.
[0106] Fig.37 A cross-sectional view of another example high voltage device is schematically shown, the device including a DFE-PD vertically isolated from a substrate and electrically connected to one or more MOS transistors and a BJT through conductive lines.
[0107] Figures 38A-38B The equivalent circuit of DFE-PD is shown. DETAILED DESCRIPTION
[0108] Introduction: Overview of BDCMOS Process Architecture Platform
[0109] Different types of semiconductor devices with different characteristics and uses include bipolar, complementary metal oxide (CMOS), and double diffused MOS (DMOS) devices. For example, bipolar devices (such as bipolar transistors) can provide accurate and high-speed analog functions, CMOS devices can provide high-speed and high-density digital functions, and DMOS devices can support high-power and high-voltage operations. Traditional high-performance bipolar and CMOS devices operate at relatively low voltages (<5 volts), while DMOS devices can operate at high voltages (up to 400 volts). Many applications can benefit from the functionality provided by a monolithically fabricated circuit or device that includes a combination of these devices integrated and interconnected on a single substrate (or chip). An integrated circuit (IC) device that includes a combination of monolithically fabricated bipolar junction devices and CMOS devices can be referred to as a BiCMOS device or platform. In some cases, a BiCMOS device or platform that includes MOS and DMOS transistors or low-voltage and high-voltage CMOS devices can be referred to as a bipolar CMOS-DMOS (BCDMOS) device or platform.
[0110] However, because bipolar, CMOS, and DMOS devices employ different structural features and design rules and may need to operate at different voltage levels, it may be relatively difficult, inefficient, and expensive to manufacture (e.g., monolithically) BiCMOS or BCDMOS devices on a common substrate, or to manufacture a monolithic circuit with BCDMOS devices on a single chip. For example, when CMOS manufacturing technology is used to co-manufacture CMOS transistors (also known as MOS transistors) and BJTs with conventional designs, one or both of the BJTs and MOS transistors may lose technical and / or cost competitiveness to be co-manufactured on the same platform and using the same or substantially the same manufacturing process.
[0111] To avoid the complexities associated with monolithic BiCMOS or BCDMOS platforms, some conventional BCDMOS devices are formed by connecting bipolar, CMOS, and DMOS devices fabricated on separate chips (or dies). Figure 1A An example conventional method of manufacturing BCDMOS devices is shown, which includes a multi-die solution in which different devices are manufactured on separate substrates and combined into a single package or multi-die platform 100. In many existing manufacturing methods, bipolar devices, CMOS devices, and DMOS devices are manufactured on separate dies using different manufacturing technologies and then mounted (e.g., surface bonding or soldering) on a common substrate 101. However, this approach may have many disadvantages, including high cost, high packaging, high complexity, inefficient area integration, multi-site manufacturing, and sometimes low reliability complexity, to name a few. On the other hand, because these different devices may employ very different physical design rules and thermal budgets and other differences, integrating these devices on the same substrate may involve little synergy and result in unacceptably long manufacturing times.
[0112] Manufacturing BCDMOS devices on a single chip, where different devices are manufactured on the same semiconductor substrate (e.g., monolithic manufacturing), can provide advantages such as reduced area occupation, lower manufacturing costs, and large-scale integration if substantial process steps can be coordinated between different devices in bipolar devices, CMOS devices, and DMOS devices. Figure 1B1 is a block diagram illustrating the concept of a monolithically fabricated analog or mixed analog and digital platform (referred to herein as a superset 102) that includes at least three transistors with different functions (e.g., high gain, fast switching, high voltage operation, fast power and / or voltage control, etc.). The superset 102 can control electronic signals (e.g., digital and analog signals) over a wide voltage and power range. In some implementations, the transistor technology may include DMOS, CMOS, and complementary bipolar (C-Bipolar) transistors. The design concept of the superset 102 combines various unique features and advantages of different types of devices, including the high switching speed and density (number of devices per unit area) of CMOS transistors (e.g., field effect transistors or FETs), the fast analog power control supported by bipolar junction transistors (BJTs), and the high power and high voltage handling capabilities of DMOS transistors. For example, CMOS transistors can operate between 3-125 volts, DMOS transistors can operate between 12-225 volts, and C-bipolar transistors can operate between 3-36 volts. However, since different devices, including CMOS, DMOS, and bipolar devices, are manufactured using very different process technologies, this strategy can still be inefficient in terms of time and cost due to high lithography mask counts and slow factory cycle times. In addition, a process architecture and process flow designed for one application, such as with a unique voltage node, may be very unsuitable for other applications.
[0113] In principle, the superset 102 can be manufactured using a currently available or commercialized BCDMOS manufacturing platform or a combination of BCDMOS and JFET manufacturing platforms. However, the manufacturing of such a superset based on these two technologies may be affected by disadvantages such as high manufacturing cost, long cycle time, high mask count, and lithography limitations. Although combining the existing JFET platform with the BCDMOS platform can improve the performance of devices (including bipolar devices) used in the superset 102, it will inevitably increase the disadvantages such as cost, cycle time, and mask count unfavorably compared to BCDMOS.
[0114] While combining DMOS devices with bipolar devices on a single platform enables high voltage (e.g., >10 volts) operation of BCDMOS devices, many applications can benefit from or may require high voltage operation of bipolar devices (e.g., BJTs that can handle voltages above 10 volts, 100 volts, or even 200 volts). However, the trade-offs between device characteristics associated with high voltage, high gain (e.g., >100), high speed (e.g., >1 GHz), and linearity (among other parameters) impose limitations on transistor designs (e.g., BJT designs) that can control high voltage signals at high speed. Furthermore, modifying low voltage devices for high voltage operation while keeping their overall size within acceptable limits for large-scale integration can be very challenging.
[0115] Recognizing these and other challenges of existing BDCMOS manufacturing technology, the disclosed technology provides efficient and synergistic monolithic integration of high-performance and high-voltage BCDMOS devices, and voltage dividers that can be combined with low-voltage devices to process and control low-voltage and high-voltage signals on a single monolithically fabricated circuit. Some of the inventive devices and design methods described herein include, but are not limited to:
[0116] Low-voltage (LV) lateral (or lateral) bipolar devices (such as BJTs) that can be fabricated using mature tools, techniques, and concepts developed for CMOS-based devices, allowing them to be fabricated monolithically with CMOS devices (such as MOSFETs).
[0117] Design concepts that can increase the operating voltage of low-voltage lateral bipolar devices using existing techniques and concepts developed for DCMOS-based devices and other approaches.
[0118] Physically partition the low-voltage (5V or lower) / high-performance region and the high-voltage (5-400V) region of the BCDMOS process architecture to design BCDMOS devices with a high-voltage region that is physically and electrically isolated from the low-voltage region.
[0119] Depletion field effect (DFE) voltage dividers (PDs) that can be connected in series with one or more low voltages to create a high voltage device. DFE PDs can be fabricated using mature tools, techniques, and concepts developed for CMOS-based devices, allowing them to be fabricated monolithically with BCDMOS devices.
[0120] On-chip isolation structures and designs for electrically isolating high-voltage and low-voltage devices monolithically fabricated on the same substrate.
[0121] Various inventive aspects disclosed herein may include the purposeful construction of a bipolar device (e.g., a lateral BJT) in which at least one feature of the bipolar device has substantially the same physical dimensions as a corresponding feature of a MOS device, for example, because the corresponding features are co-fabricated using masks or process steps developed for the same or similar technology nodes or design rules.
[0122] In various conventional BiCMOS processes, BJTs are fabricated vertically to allow control of various device parameters, such as vertical doping profiles and base dimensions (e.g., width and / or thickness of the base region). This approach requires separate fabrication steps, as well as masks for fabricating BJTs and MOS transistors on a common substrate, and significantly increases the cost, complexity, and cycle time of IC devices.
[0123] A first inventive aspect of a BiCMOS core process architecture is disclosed herein that facilitates the construction of a high-performance, low-cost, low-voltage, and fast factory cycle-time BiCMOS technology platform using existing digital cell libraries and manufacturing processes established for CMOS devices. Some device designs described herein (e.g., lateral bipolar device designs) can synergistically benefit from manufacturing steps used in CMOS manufacturing technologies that use the same components and elements. Therefore, the bipolar device designs disclosed herein can utilize the precision of CMOS processes that have been developed for decades to manufacture high-performance bipolar devices (e.g., bipolar transistors). In some embodiments, the disclosed bipolar devices (e.g., lateral bipolar junction transistors) can be co-manufactured with CMOS devices (e.g., MOS field effect transistors) through co-manufacturing or parallel manufacturing process flows, wherein one or more manufacturing steps can be at least partially simultaneously or simultaneously shared. In some embodiments, the disclosed bipolar devices (e.g., lateral BJTs) can be manufactured substantially completely independently of MOS devices, but can still benefit from the precision, low cost, and shortened cycle time provided by developed and advanced CMOS manufacturing technologies.
[0124] An example of a BiCMOS platform is an IC device that includes a metal oxide semiconductor (MOS) transistor (e.g., a MOS field effect transistor or MOSFET) and a bipolar junction transistor (BJT), the metal oxide semiconductor (MOS) transistor including a gate stack formed on its channel region and the bipolar junction transistor (BJT) including a layer stack formed on its collector region. The MOS and bipolar transistors of the disclosed IC device are constructed using some common CMOS process features. Therefore, various features of the bipolar device (e.g., BJT) are defined using structural features and process steps used to manufacture MOS devices (e.g., MOSFET). In some embodiments, implants of lightly doped and / or highly doped regions used to form source and drain regions, gate regions, and channel regions of a CMOS device (e.g., MOSFET) are used to form emitter, base, and collector regions of a bipolar device (e.g., BJT). As described herein, highly doped or heavily doped (HD) regions, sometimes denoted as P ++ Area or N ++ area, can have more than about 1×10 19 cm -3 , more than about 1×10 20 cm -3 A peak doping concentration of , or a value within a range defined by any of these values. In addition, a lightly doped (LD) region, sometimes denoted N - or P - region, where the peak doping concentration may be less than about 1×10 14 cm -3 or about 1×1013 cm -3 In some cases, highly doped or heavily doped (HD) regions (e.g. P ++ or N ++ ) may include an ohmic contact region that provides electrical connection to a region (e.g., a region forming a HD region). For example, a spacer layer formed on the sidewall of a MOSFET gate stack is used to define a lightly doped (LD) drain (LDD) region, which is used to define a base length (L) of a BJT. B ). In addition, the gate contact layer (e.g., silicide layer) and gate layer (e.g., polysilicon layer) forming the gate stack of the CMOS device are used to form a field plate of a bipolar device, such as a reduced surface field (RESURF) plate (e.g., for increasing its operating voltage). As described below, one or both of the spacer layer and the field plate can also be advantageously used as an implant hard mask.
[0125] An example of a disclosed bipolar device that can be co-fabricated with a CMOS device is a lateral BJT (L-BJT) having an emitter region, a base region, and a collector region arranged in a lateral direction parallel to a major surface of a common substrate on which a MOS transistor (e.g., a MOSFET) is fabricated. The L-BJT can include a PNP (or NPN) transistor having a laterally formed PN (or NP) and NP (or PN) junction configured to support bipolar transistor operation based on bipolar carrier transport substantially in a lateral direction (the path of the bipolar current is substantially in a lateral direction). The L-BJT arrangement produces several corresponding features between the two types of devices and allows for enhanced parallel processing of various manufacturing steps between MOSFETs and BJTs. Therefore, replacing a conventional vertical BJT design with an L-BJT can reduce cycle time by significantly improving the synergy between the CMOS and BJT manufacturing process flows.
[0126] Figure 2 Some examples of correspondence between features are shown in . Figure 2A portion of an integrated circuit (IC) device 200 fabricated using a low voltage, high performance BiCMOS architecture and process flow according to an embodiment is shown. According to an embodiment, the IC device 200 may include a PNP lateral bipolar junction transistor (PNPL-BJT) 240 and an n-channel MOS (NMOS) transistor 280 or both fabricated according to a low voltage, high performance BiCMOS process architecture. The IC 200 may also include one or both of a p-channel MOS (PMOS) transistor 230 and an NPN L-BJT 235. The PMOS transistor 230 has similar features to the NMOS transistor 280, but its region includes a semiconductor region of opposite polarity type compared to a corresponding region of the n-channel NMOS transistor 280. Similarly, the NPN L-BJT 235 has similar features to the PNP L-BJT 240, but its region includes a semiconductor region of opposite polarity type compared to a corresponding region of the PNP L-BJT 240. The well (eg, common well 212 ) in which the MOS transistors 230 , 280 and the L-BJTs 240 and 235 are formed may be produced in a substrate 290 (eg, a silicon substrate).
[0127] It should be understood that, as described herein, while some features may be described with reference to a particular type of MOS transistor (e.g., one of the NMOS transistor 280 or the PMOS transistor 235), it should be understood that, where applicable, similar features may exist in an opposite type of MOS transistor (e.g., the other of the PMOS transistor 235 or the NMOS transistor 280). Similarly, while some features may be described with reference to a particular type of BJT (e.g., one of the PNP L-BJT 240 or the NPN L-BJT 235), it should be understood that, where applicable, similar features may exist in an opposite type of BJT (e.g., the other of the PNP L-BJT 240 or the NPN L-BJT 235).
[0128] In some embodiments, the NMOS transistor 280 may be a field effect (FET) transistor including a drain region 209, a source region 207, and a gate stack 202a formed on a channel region 203 between the drain region 209 and the source region 207. The gate stack 202a includes a gate dielectric and a gate. In addition, the NMOS transistor 280 may include a gate spacer 216a disposed on opposite sidewalls of the gate stack 202a.
[0129] In some embodiments, the PNP L-BJT 240 may include an emitter region 206, a base region 201, a collector region 211, and a layer stack 202b formed on a portion of the collector region 211. The emitter region 206 has a first polarity or semiconductor type (e.g., p-type) and is formed in a base well 208 having a second polarity or semiconductor type (e.g., n-type) opposite to the first polarity or semiconductor type.
[0130] According to various embodiments described herein, the source region and the drain region (e.g., the source region 207 and the drain region 209) of the MOS transistor (e.g., the NMOS transistor 280) may include heavily doped regions. In addition, the emitter region and the collector region (e.g., the emitter region 206 and the collector region 211) of the L-BJT may be heavily doped regions.
[0131] It should be understood that as used herein, the polarity of a region (e.g., a semiconductor region) is determined by the polarity of the majority charge carriers in the region, which in turn can be defined by the net doping. Thus, in a p-type region, the net doping is positive, such that the majority charge carriers are holes, while in an n-type region, the net doping is negative, such that the majority of charge carriers are electrons. The majority carrier in a region is a charge carrier whose concentration is greater than the concentration of the opposing charge carriers in the region, or greater than its intrinsic concentration in the absence of doping and at the same temperature. Thus, a p-type region can have a p-type dopant and an n-type dopant, but the concentration of the p-type dopant is greater than the n-type dopant, resulting in a greater concentration of holes than electrons in the region.
[0132] In some cases, emitter region 206 can have a higher doping concentration than base well 208 and base region 201. In addition, PNP L-BJT 240 can include spacer layers 216b disposed on sidewalls of layer stack 202b. In some embodiments, layer stack 202b can serve as a reduced surface field (RESURF) plate that reduces the lateral component of the electric field in drift region 205 of collector region 211. Layer stack 202b can include a dielectric layer on a substrate and a conductive layer on the dielectric layer. Emitter region 206, base region 201, and collector region 211 are arranged laterally (e.g., along the y-axis) such that the bipolar operation of PNP L-BJT 240 is primarily associated with carrier transport in the lateral direction. Therefore, base region 201 may include a region of base well 208 that extends laterally from collector region 211 to emitter region 206, with a base length (L B) can be defined by the lateral distance from the vertical interface between the base well 201 and the collector region 211 (e.g., along the z-axis) to the vertical interface between the emitter region 206 and the base well 208 closer to the collector region 211 (referred to herein as a vertical emitter-base junction). Advantageously, in some embodiments, the base length (L) can significantly affect the performance of the PNP L-BJT 240. B ) is defined by the lateral width of the bottom or base portion of the spacer 216b (e.g., the spacer or spacer portion above the base region 201), which in turn is defined by the spacer manufacturing process. Assuming that the dimensions of the spacer 216b (and correspondingly the spacer 216a) can be precisely defined using standard CMOS manufacturing processes, the base length of the PNP L-BJT 240 can be precisely controlled (or defined) by a common manufacturing process. Thus, the design and structure of the L-BJT not only enables it to be co-fabricated with MOS transistors on a common substrate, but also enables its base length to be precisely controlled (e.g., with a resolution of ±5nm, ±3nm, ±1nm, or sun nanometers). This level of control over the base length approaches that of vertical BJTs used in conventional BiCMOS platforms, which require dedicated manufacturing processes that cannot be synergistically combined with MOS transistor manufacturing processes. In some cases, the base length can be defined by the lateral width of the bottom of the spacer. This space can be formed by depositing a polysilicon layer and vertically etching the polysilicon layer to provide the desired bottom thickness. Thus, by controlling the polysilicon thickness (by controlling its growth time) and the etching time, the width of the bottom of the spacer layer (and thus the base length) can be precisely defined.
[0133] In some examples, the entire structure of the L-BJT (the base well and emitter region, the layer stack, and the collector region therein) can be fabricated in and / or on the common well. In some other examples, such as the PNP L-BJT 240, the L-BJT can be partially formed in the common well 212 and the second region (e.g., the second well) 214, the second region 214 including a semiconductor having an opposite dopant type and / or an opposite polarity compared to the common well 212. In these examples, the common well 212 and the second region 214 can have an interface below the base well 208, such that a first portion of the base well 208 shares an interface with the common well 210, and a second portion of the base well 208 shares an interface with the second well. In some cases, the vertical interface between the common well 212 and the second region 214 can be aligned with the vertical emitter-base junction of the PNP L-BJT 240. In some cases, PNP L-BJT 240 may include a highly doped (HD) base region (not shown) disposed on base well 208 to provide an ohmic contact with base well 208 and base region 201 therein.
[0134] Still refer to Figure 2 In the example shown, the NMOS transistor 280 and the MOS transistor 230 are fabricated in the common well 212 and the second well 214, respectively. The PNP L-BJT 240 is partially fabricated in the common well 212 and the second well 214, and the NPN L-BJT 235 is fabricated in the third well 215 and the common well 212. The second well 214 and the third well 215 have the same polarity (e.g., n-type), and the common well 212 has a polarity opposite to that of the second well 214 and the third well 215.
[0135] According to various aspects, the MOS transistors 280, 230 and the L-BJTs 240, 235 are cooperatively formed using various common structural features. As a result, the gate stack 202a of the NMOS transistor 280 and the layer stack 202b of the PNP L-BJT 240 have one or more corresponding layers having common physical dimensions or materials. For example, one or both of the dielectric layer and the conductive layer of the layer stack 202b can be formed simultaneously or using the same process recipe, and thus can be formed of the gate dielectric and the gate of the gate stack 202a and have the same thickness. In addition, in some embodiments, the gate stack 202a and the layer stack 202b can be patterned using a mask designed with similar design rules, so that one or more lateral dimensions of the gate stack 202a and the layer stack 202b can be substantially the same. In addition, the gate stack 202a of the MOS 280 transistor and the layer stack 202b of the PNP L-BJT 240 have corresponding spacer layer structures 216a, 216b formed on their sidewalls having common physical dimensions. Similar to the layer stack 202b, the spacer layer structures 216a, 216b can be formed at the same time or using the same process recipe, and thus are formed of the same material and have the same physical dimensions. In addition, the NMOS transistor 280 and the PNP L-BJT 240 have corresponding implant diffusion regions with a common implant dopant distribution, for example, due to being formed at the same time or using the same implant mask and / or recipe, and thus have the same implant distribution in one or more directions.
[0136] Advantageously, the lateral arrangement of emitter region 206 , base region 201 , and collector region 211 in a direction parallel to a major surface of common substrate 290 allows for enhanced simultaneous or parallel processing and / or sharing of process recipes between various steps between MOS transistors 280 , 230 and L-BJTs 240 , 235 .
[0137] For example, patterning including photolithography and etching for defining the gate stack 202a of the NMOS transistor 280 and the layer stack 202b of the PNP L-BJT 240 can be processed together, i.e., processed simultaneously. As a result, according to an embodiment, the layer stack 202b on the collector region 211 of the PNP L-BJT 240 can have a dielectric layer co-deposited or co-formed by oxidation with the gate dielectric of the NMOS transistor 280, and can have the same thickness as the gate dielectric. Similarly, the layer stack 202b of the PNP L-BJT 240 has a conductive layer on the dielectric, which is co-deposited with the gate electrode on the gate dielectric of the NMOS transistor 280 and can have the same thickness. After jointly forming the layer stack 202 b of the PNP L-BJT 240 and the gate stack 202 a of the NMOS transistor 280, the layer stack 202 b of the PNP L-BJT 240 and the gate stack 202 a of the NMOS transistor 280 may be jointly patterned using a shared photolithography and etching process. In some cases, the layer stack 202 b of the PNP L-BJT 240 and the gate stack 202 a of the NMOS transistor 280 may correspond to the length of the drift region 205 of the BJT and the channel length of the MOS transistor in the first direction, respectively.
[0138] Thus, the patterned or co-patterned gate stack 202b and layer stack 202b can be used as a hard mask for ion implantation. For example, the corresponding regions can be co-implanted using the same energy and / or dose, and can also have substantially the same dopant distribution. For example, using the gate stack 202b and layer stack 202b as a patterned layer, which can also include a patterned layer left thereon, such as a photoresist or a hard mask, the lightly doped (LD) source and drain regions of the n-channel NMOS transistor 280 can be co-implanted with the base well 208 of the PNP L-BJT 240. Similarly, the LD source and drain regions of the p-channel MOS transistor 230 can be co-implanted with the base well of the NPN L-BJT 235.
[0139] After the LD implant, spacer layers 216a, 216b may be co-formed on the sidewalls of the gate stack 202a of the NMOS transistor 280 and the layer stack of the PNP L-BJT 240. As a result, the spacer layer 216b of the PNP L-BJT 240 and the spacer layer 216a of the NMOS transistor 280 may be formed by co-depositing spacer layers (e.g., nitride or oxide layers) and co-etching (e.g., anisotropic co-etching) to have the same lateral dimensions and define one or more implant diffusion regions for the PNP L-BJT 240 and the NMOS transistor 280 having the same type of implant (and in some cases, a common implant dopant profile).
[0140] Using the spacer layer 216b of the PNP L-BJT 240 and the spacer layer 216a of the NMOS transistor 280 that have been formed together, the emitter region 206 of the PNP L-BJT 240 and the source and drain (S / D) regions 207, 209 of the MOS transistor can be defined by co-implantation. The NMOS transistor 280 is an n-channel MOSFET and the PNP L-BJT 240 is a PNP BJT, so the S / D regions are 207, 209, and the emitter region 206 can be n-doped, such as co-implanted with an n-type dopant. Similarly, the emitter regions of the S / D region MOS transistor 230 (which is a p-channel MOSFET) and the NPN L-BJT 235 (which is an NPN BJT) are p-doped, such as co-implanted with a p-type dopant.
[0141] In addition, before forming a gate stack (e.g., gate stack 202a) and a layer stack (e.g., layer stack 202b), a common patterned layer can be used to define the collector region of the BJT and the channel region of the MOS transistor by a co-implantation process. When the channel region is a p-doped well region (e.g., channel region 203) of an n-channel MOS transistor (e.g., NMOS transistor 280), it can be co-implanted with a p-type collector region (e.g., collector region 211) of a PNPL-BJT (e.g., PNP L-BJT 240). Similarly, when the channel region is an n-doped well region of a p-channel MOS transistor (e.g., PMOS transistor 230), it can be co-implanted with an n-type collector region of an NPN L-BJT (e.g., NPN L-BJT 235).
[0142] Other regions of the highly doped (HD) and lightly doped (LD) collector regions of the L-BJT may be defined by co-implanting with corresponding HD and LD implants in the S / D regions of the MOS transistors. For a PNP L-BJT (e.g., PNP L-BJT 240), the p-type collector region (e.g., collector region 211) includes an HD region co-implanted with a p-doped S / D region of a p-channel MOS transistor (e.g., MOS transistor 230) and a p-type emitter region 206 of the PNP. For an NPN L-BJT (e.g., NPN L-BJT 235), the n-type collector region includes an HD region co-implanted with an n-doped S / D region of an n-channel MOS transistor (e.g., NMOS transistor 280) and an n-type emitter region of the NPN L-BJT 235.
[0143] For a PNP L-BJT (e.g., PNP L-BJT 240), the p-type collector region 211 includes an LD doped region co-implanted with a p-doped LD S / D region, a p-channel MOS transistor (e.g., MOS transistor 230), and a p-type base well of an NPN L-BJT (e.g., NPN L-BJT 235). Similarly, for an NPN L-BJT (e.g., NPN L-BJT 235), the n-type collector region includes an LD doped region co-implanted with an n-doped LD S / D region of an n-channel MOS transistor (e.g., NMOS transistor 280) and an n-type base well (e.g., base well 208 and base region 201 therein) of a PN L-BJT (e.g., PNP L-BJT 240).
[0144] In the above, various corresponding structures of CMOS and L-BJT devices are described as being advantageously suitable for co-manufacturing, such as simultaneous manufacturing, according to embodiments. However, embodiments are not limited thereto, and corresponding structures can be manufactured sequentially. Even in the case of sequential manufacturing, it will be appreciated that various process recipes, such as deposition, etching, and implantation recipes, can be used in synergy to manufacture corresponding structures in CMOS and L-BJT devices due to shared design rules.
[0145] MOS transistor and L-BJT with extended drain and drift region
[0146] As described above, the above designs and structures can be used to co-fabricate L-BJTs and MOS transistors (MOSFETs) for use in a BiCMOS platform. In some cases, these L-BJTs may be low voltage (LV) BJTs with a maximum operating voltage of less than 5 volts. Therefore, the operating voltage of BiCMOS circuits fabricated based on conventional MOS transistors (e.g., MOSFETs) and the disclosed L-BJTs may be limited, while for some applications, MOSFETs and BJTs operating at higher voltages (e.g., greater than 10 volts, 100 volts, 200 volts, or 300 volts) and co-fabricated on a common substrate may be required.
[0147] To meet these and other needs, the second inventive aspect described herein is intended to increase the operating voltage of a bipolar junction device (e.g., a lateral BJT) fabricated together with a MOS transistor on a common substrate, for example by increasing the length of at least one drift region of the bipolar junction device. Similarly, by increasing the length of the channel region of the MOS transistor, the operating voltage of the MOS transistor (e.g., a MOSFET) can be raised to a higher voltage range (e.g., 5-400V). In some embodiments, the drift region of the L-BJT can extend from the LD collector region to the base region of the L-BJT in a lateral direction parallel to the common substrate. Similarly, the channel region of the MOS transistor can extend from the LD source region to the LD drain region in a lateral direction. In some cases, by keeping the magnitude of the lateral component of the electric field along the drift region or the channel region constant, including an extended gate stack on the channel region of the MOSFET and / or including an extension layer stack on the drift region of the L-BJT can further increase the operating voltage of the device.
[0148] Figure 3A A side cross-sectional view of a high voltage (HV) MOSFET 300 having an extended drain region and an extended gate stack is schematically shown. The extended drain region includes a channel region 203 and an extended drain region 304a. The extended gate stack includes a gate stack 202a and a gate stack extension 302a. Similar to MOSFET 280, HV MOSFET 300 includes a source region 207 and a drain region 309. In some examples, the drain region 309 of HV MOSFET 300 includes a highly doped region, while the drain region 209 of MOSFET 280 may include highly doped and lightly doped regions.
[0149] Figure 3B A side cross-sectional view of a high voltage (HV) L-BJT 350 having an extended drift region and an extended layer stack is schematically shown. The extended drift region includes a drift region 205 and a drift region extension 304b. The extended layer stack includes a layer stack 202b and a layer stack extension 302b. Similar to the PNP L-BJT 240, the HV L-BJT 350 includes an emitter region 206 and a base region 201 located at one end of the extended drift region, and a collector region 311 located closer to the other end of the layer stack extension 302b when the drift region is extended. In some examples, the collector region 311 of the HV L-BJT 350 includes a highly doped region, while the collector region 211 of the MOSFET 240 may include highly doped and lightly doped regions.
[0150] In some embodiments, HV L-BJT 350 is co-fabricated with MOSFET 300 on a common substrate 290. Advantageously, the longer drift, channel, layer stack, and gate stack of HV MOSFET 300 and HV BJT 350 enable their maximum operating voltage to be greater than the corresponding MOSFET 280 and PNP L-BJT 240. A voltage applied to the HD drain region of HV MOSFET 300 or the HD collector region of HV L-BJT 350 may drop along the extended drain region and drift region, respectively. In some cases, the voltage may drop by at least 50% from the HD drain region of HV MOSFET 300 and the HD collector region of HV L-BJT 350 to the source region 209 and the base region 201, respectively.
[0151] As used herein, the maximum operating voltage (also referred to as the breakdown voltage) of a MOS device (e.g., MOSFET) or a bipolar device (e.g., L-BJT) may be the voltage at which at least one junction and / or at least one dielectric layer within the device electrically breaks down under reverse bias.
[0152] HV MOSFET 300 and HV L-BJT 350 may include one or more features described above with respect to NMOS transistor 280 and PNP L-BJT 240. Thus, in some examples, one or more regions and / or layers of HV MOSFET 300 may be co-fabricated with corresponding regions and / or layers of HV L-BJT 350 and have the same or substantially the same physical dimensions. In addition, one or both of the extended layer stack of HV L-BJT 350 and the extended gate stack of HV MOSFET 300 may have the same physical dimensions and may be co-fabricated in the same manufacturing steps.
[0153] In some embodiments, at least one structure, one dimension, or one doping profile of the above-described L-BJT design may be different from the structure, dimension, or doping profile of a corresponding feature in a MOS transistor fabricated with the L-BJT. As a result, at least one fabrication step of the L-BJT may be a dedicated fabrication step that is unrelated to the fabrication of the corresponding feature in the MOS transistor. The dedicated fabrication steps may include, but are not limited to, forming or doping regions (e.g., diffusing dopants), etching structures (e.g., spacer layers), setting layers, etc.
[0154] Voltage Divider
[0155] A third inventive aspect of the disclosed technology is to boost the low voltage BiCMOS platform to a higher voltage node by using a high voltage (HV) divider. The inventive aspect arises from the inventors' recognition that high voltage devices based on the BCDMOS process platform, such as the MOS and bipolar devices described above with extended drift and channel regions, can be conceptually divided into two regions. In some cases, a first region of the device is designed to operate at a low voltage (e.g., less than 5 volts), and a second region of the device can be physically separated or at least partially isolated from the first region and designed to provide a voltage drop from a higher voltage (e.g., 5-400 volts) to a safe voltage level that can be applied to the first region. In some examples, the first region can include a high performance and low voltage channel portion of a MOS device (e.g., a MOSFET), or a high function and low voltage emitter base region of a bipolar device (e.g., an L-BJT). The second region can include an extended drift region or an extended drain region that converts a high voltage provided to a HD drain region or HD collector region to a low voltage near the source region or base region of a MOS transistor or LBJT, respectively.
[0156] Recognizing this conceptual division between the low voltage / high performance (5V or lower) region and the voltage drop region of the BCDMOS process architecture, aspects of the present invention include a BCDMOS process architecture that can support a wide range of voltages by physically dividing a MOS or bipolar device into a first region that supports high performance at low voltage and a second region that includes a voltage divider (PD) that can be electrically connected or connected in series with any of the low voltage regions or components to create a high voltage device. Both regions can be fabricated on the same substrate using a low cost and fast cycle time bipolar CMOS (BiCMOS) core process and platform, examples of which have been described above.
[0157] Figure 4A Schematically shows a side cross-sectional view of an HV MOSFET 300 with an extended drain region. A first region 404 of the HV MOSFET 300 may be a low voltage MOS portion including a source region 207 and at least a portion of a gate stack 202a. A second region 402 of the HV MOSFET 300 may be a PD portion including an extended drain region and a drain region 209. Figure 4B A side cross-sectional view of a lateral bipolar junction transistor (L-BJT) 350 with an extended drift region is schematically shown. A first region 454 of the HV L-BJT 350 may be a low voltage bipolar portion including emitter and base regions and at least a portion of a layer stack. A second region 452 of the HV L-BJT 350 may be a PD portion including an extended drift region and a collector region.
[0158] Although the operating voltage of a MOS or bipolar device can be increased (e.g., above 5 volts) by increasing the length of the channel or drift region, applying a high voltage to a region of the integrated device that is physically close to a region designed to handle lower voltage levels (e.g., the first region 404 or 454) may reduce the reliability of the device. For example, certain parasitic effects may couple a portion of the high voltage to the first region 404 (or first region 454) of the HV MOSFET 300 (or L-BJT 350) and cause junction breakdown or dielectric breakdown in the first region 404.
[0159] In some embodiments, physically separating the first region 404, 454 (low voltage portion) from the corresponding second region 402, 452 (voltage divider portion) can improve the reliability of the corresponding device, for example, by reducing the breakdown probability caused by parasitic voltage coupling between the low voltage and the voltage divider. In addition, physically separating the first region 404, 454 (low voltage portion) from the corresponding second region 402, 452 (voltage divider portion) can prevent minority carriers generated by weak impact ionization from attacking the delicate gate oxide. In addition, the physical separation between the low voltage and voltage divider portions of MOS and bipolar devices can allow the use of a voltage divider to couple multiple devices of the same or different types. For example, a voltage divider can receive a high voltage and provide a reduced voltage to two or more devices including an L-BJT and a MOSFET.
[0160] In various embodiments, physical separation can include reducing / or limiting electrical, magnetic or electromagnetic coupling between the first and second regions by isolation (referred to herein as isolation). Therefore, physical separation can include providing a distance and / or isolation element, layer or structure between the low voltage portion and the PD portion of the device. The distance may include a lateral distance, a vertical distance or a combination thereof. In some cases, physical isolation that provides sufficient isolation to reduce the risk of breakdown of the low voltage portion may be converted into a distance that significantly increases the overall size of the device or the corresponding circuit. Therefore, in some embodiments, one or more isolation elements, layers or structures can be used to isolate one or more low voltage device portions or low voltage substrate regions from one or more voltage divider portions.
[0161] In some cases, the PD portion may include a substrate region (referred to herein as a PD substrate region) that is physically separated from a substrate region that includes an L-BJT and / or a MOSFET. The L-BJT includes a collector region that is electrically connected in series to the PD substrate region and physically separated therefrom by an isolation structure. In some embodiments, the L-BJT also includes a layer stack that includes a conductive field plate formed on the collector region. In some other embodiments, the collector region and the PD substrate region are electrically connected by one or more metallization layers formed above the main surface of a common semiconductor substrate. In some other embodiments, the PD substrate region is configured to reduce more than 50% of the voltage applied across the BJT and the PD substrate region. According to an embodiment, the PD substrate region formed in the same common substrate but physically separated from the LV region including the L-BJT and the CMOS device can be isolated from the LV region by one or more isolation structures, for example, a dielectric isolation structure and / or a junction or well isolation structure including a deep trench or a shallow trench filled with a dielectric material (e.g., an oxide such as silicon dioxide).
[0162] Still refer to Figure 4A and 4B In some embodiments, the PD portion (second region 402, 452) described above can be designed and optimized as an independent PD device in the context of a BJT or MOSFET with an extended drain and drift region. Connecting a low voltage (LV) device or platform in series with such a PD device can increase the operating voltage of the platform without changing the device or platform structure. In some cases, the PD device can be physically separated from the platform or LV device based on the principles described above regarding the physical separation between the first and second regions of the HV MOSFET 300 and the HV BJT 350, or the physical separation of the PD substrate region from the LV substrate region.
[0163] In some cases, the platform or device may include an integrated circuit (IC) device. According to various aspects, the IC device includes a metal oxide semiconductor (MOS) transistor and a bipolar junction transistor (BJT) formed in a common semiconductor substrate. The PD device can be connected in series with one or more low voltage devices to provide a variety of high voltage MOS and bipolar devices (e.g., HVDMOS, HV BJT, HV MOS, etc.). This enables the BiCMOS core process technology to adapt to a wide range of application voltage nodes with relatively little or no additional cost, or to redesign the various device structures used to manufacture these devices.
[0164] A fourth inventive aspect is the development and optimization of a general PD device that can be manufactured using a CMOS process. In various embodiments, the PD device may include an input port, region, or electrode that receives a high voltage (e.g., 5-400V), outputs a low voltage (e.g., 0-5V) and is configured to be connected to an output port, region, or electrode of an LV device. Thus, the PD device is connected in series between the device providing the high voltage signal and the LV device. In some examples, the PD includes one or more RESURF electrodes or connections that are configured to be connected to a ground potential or a stable low potential of a device that provides a reduced voltage to the LV platform or PD.
[0165] Figures 5A-5C Example embodiments for MOS and bipolar junction devices connected to a voltage divider (PD) 500 are shown. Figure 5A There is schematically shown an isolated PD 500 connected to an NMOS transistor 280. A RESURF electrode of the PD 500 is electrically connected to a gate stack of the NMOS transistor 280, and an output port of the PD 500 is electrically connected to a HD drain region of the NMOS transistor 280.
[0166] Figure 5B Schematically illustrated is an isolated PD 500 connected to a vertical bipolar junction transistor (V-BJT) 506. The output port of PD 500 is electrically connected to the collector region of V-BJT 506. The RESURF electrode of PD 500 is electrically connected to the highest relative potential, i.e., for an N-type device, a positive voltage will be applied to the PD, and the RESURF electrode is connected to the lowest potential (e.g., ground potential). In some cases, the RESURF electrode of PD 500 can be connected to an independent voltage source.
[0167] Figure 5C Schematically illustrated is an isolated PD 500 connected to a lateral bipolar junction transistor (L-BJT) 240. The RESURF electrode of the PD 500 is electrically connected to the layer stack of the PNP L-BJT 240, and the output port of the PD 500 is electrically connected to the HD collector region of the PNP L-BJT 240.
[0168] In various embodiments, the electrical connection 502 between the RESURF electrode of PD 500 and the gate stack, layer stack of the LV device, and the electrical connection 504 between the output port of PD 500 and the collector or drain region of the LV device may include connections through wires in the metallization layer of an IC that includes PD 500 and the corresponding LV device. For example, two or more conductive vias may be used to connect the RESURF electrode and the output port of PD 500 and the corresponding region of the LV device to a common wire in the metallization layer above PD 500 and the LV device.
[0169] When the LV device is an L-BJT ( Figure 5C ), various coordinated parallel processing with MOS transistors can be performed, including defining the gate stack of the MOS transistor and the layer stack of the BJT, defining the spacer layer of the MOS transistor and the BJT and the related implantation diffusion and defining various implantation diffusion regions. The details of the coordinated processing have been described above and will not be repeated here for the sake of brevity.
[0170] In certain embodiments described herein, the PD device may be a depletion field effect voltage divider (DFE-PD) that monotonically (e.g., substantially linearly) drops a voltage from a higher input voltage at a first node (input node) to a lower voltage at a second node (output node). Thus, the semiconductor device is suitable for analog signal conversion. The PD includes separately doped silicon regions (e.g., four regions, or in some embodiments, five or more regions) that are arranged to form a reverse biased depletion junction (first junction) at the input node and a different junction (second junction) at the output node. Charge is coupled from the first junction to the second junction to act as a voltage divider, and the ratio of charge coupling defines the ratio of voltage scaling. In certain implementations, the full range of scaling can range from direct one-to-one voltage conversion to a situation where the output voltage is independent of the input voltage and can be fixed to a substantially constant value.
[0171] Devices using the above voltage reduction scheme can be implemented in a variety of electronic devices. Examples of electronic devices include, but are not limited to, industrial automation / control systems, military and aerospace systems, security and surveillance systems, building technology, instrumentation and measurement systems, power management systems, data acquisition systems, radio frequency communication systems, consumer electronics, electronic test equipment, communication infrastructure, radar systems, mobile devices (e.g., smartphones or mobile phones), phased array antenna systems, laptops, tablets, and / or wearable electronic devices.
[0172] Vertical Isolation of Voltage Divider
[0173] A fifth inventive aspect includes architectures, designs, and structures for isolating a PD device (e.g., a DFE-PD) or a portion of a PD from a LV device or a portion of a LV device fed by a PD device. These architectures, designs, and structures can reduce the cost, complexity, and / or cycle time of manufacturing a high voltage BCDMOS platform, facilitate connecting a PD device to multiple HV devices in the same or different IC layers, and / or enable high bandgap materials to be included in the structure of a PD device.
[0174] In some implementations, the PD device or region may be formed in the front end of line (FEOL) in the same silicon substrate according to some embodiments. In these embodiments, the BiCMOS device may be isolated from the HV divider by an isolation structure such as shallow trench isolation (STI). Figures 5A-5C As shown, such an isolation scheme may include using a lateral isolation structure to laterally isolate the PD device (or PD substrate area) from the LV device (or LV substrate area). The lateral isolation structure may include a trench filled with a dielectric material (e.g., silicon dioxide). In some cases, the trench may be a deep trench extending from a metallization layer to a common substrate forming the LV device, and the PD device or a portion thereof is connected to the LV device through the metallization layer. In some cases, the width of the trench may be determined by the amplitude of the input voltage level provided to the PD device. In some examples, multiple lateral isolation structures may be provided at different locations in the IC to isolate the PD device from different lateral directions. In some cases, the lateral distance of the lateral isolation structure from the PD device and the LV device may be determined by the width of the trench and the input voltage level provided to the PD device.
[0175] In some embodiments, for modest voltage ranges (e.g., less than 50V), a junction boundary can be used to laterally isolate a PD device or portion from a LV device or portion. In some cases, the depletion of the junction boundary (isolation junction) can be determined based on the magnitude of the voltage drop from the PD device or region input node to the LV device or region. In some cases, the depletion width of the isolation junction can be 10 to 20 microns. Where possible, using a junction boundary as an isolation structure (e.g., a lateral isolation structure) can reduce the manufacturing cost of the IC. However, given that the depletion width of the junction boundary may scale with the magnitude of the voltage drop, this isolation scheme may consume more substrate space.
[0176] In alternative implementations, the voltage divider may be formed at the back end of line (BEOL) above the substrate. In these embodiments, the BiCMOS device may be vertically isolated from the PD by at least one inter-metal or inter-layer dielectric (IMD or ILD) layer, thereby eliminating the need for lateral isolation at higher voltage nodes.
[0177] Fig. 6A Schematically illustrates a vertically isolated PD 600 that may be fabricated above a substrate that includes electronic devices electrically connected to the PD 600 to receive a reduced voltage. Similar to PD 500, PD 600 includes one or more input electrodes 604, one or more RESURF electrodes 606, and one or more output electrodes 608. A dielectric structure 610 (e.g., an oxide structure) vertically isolates PD 600 from the LV devices and the area below the dielectric structure 610. In some cases, dielectric structure 610 may be a capping dielectric layer 224 disposed on substrate 290 of a corresponding IC. Figure 6BSchematically illustrated is a voltage divider 612 formed in the back end of line (BEOL) within the capping dielectric layer 224. The PD 612 receives an input voltage (e.g., 5-400V) from the input electrode 604 and outputs a reduced voltage (0-5V) through the output electrode 608. The output electrode 608 may be connected to one or more input nodes or regions of a BCDMOS device fabricated on the substrate 290.
[0178] In some embodiments, vertical isolation of PD 612 can facilitate the use of wide bandgap materials to form one or more regions of PD 612. Advantageously, PDs including wide bandgap materials can improve the performance of PD 612, thereby providing better design tradeoffs for industry-leading performance. For example, PD devices including wide bandgap materials can reduce higher voltage levels (e.g., greater than 400V) to voltage levels usable by LV devices.
[0179] By adopting low voltage and high performance BiCMOS devices and PDs according to embodiments, a "superset" process architecture platform according to various embodiments can be implemented. The process architecture platform according to embodiments can greatly reduce the number of masks and manufacturing costs, for example, by up to about 50%.
[0180] Low-voltage, high-performance BiCMOS process architecture: Low-voltage lateral BJT
[0181] As described above, one inventive aspect of the disclosed technology relates to a process architecture that facilitates the construction of a high-performance, low-cost, low-voltage (5V or less), and fast cycle time BiCMOS technology platform using a digital cell library for manufacturing CMOS devices. In a conventional BiCMOS process, a BJT is manufactured with a vertical arrangement of emitter, base, and collector regions to allow control of the vertical doping profile of each layer, particularly the base length, which in this case is the vertical distance between the junction between the base-collector and the junction between the base and emitter of the BJT. The vertical direction is the direction perpendicular to the main surface of the common substrate on which the BJT and MOS devices are manufactured. Most of the manufacturing steps required to manufacture a vertical BJT (V-BJT) cannot be synergistically combined with the manufacturing steps of a MOS device, thereby increasing the cost, complexity, and cycle time of a conventional BiCMOS platform compared to a CMOS platform.
[0182] In some cases, a lateral BJT (L-BJT) design uses a lateral arrangement of emitter, base, and collector regions that are co-fabricated with corresponding features of a MOS device using CMOS manufacturing techniques to provide a high-performance bipolar junction device (e.g., BJT). In some embodiments, the disclosed L-BJT design includes regions with similar physical dimensions, dopants, doping profiles, doping regions, spacers, layer stacks, and other features that correspond to the same or similar features in a MOS transistor (e.g., MOSFET), thereby providing maximum synergy between CMOS and L-BJT manufacturing processes. As a result, the cycle time, cost, and complexity of manufacturing a BiCMOS platform based on L-BJTs can be significantly reduced compared to a BiCMOS platform based on V-BJTs.
[0183] Advantageously, the L-BJT design described above, and in greater detail below, provides synergy with CMOS manufacturing without sacrificing the high performance required of BJTs on the BiCMOS platform. For example, because the base length of the L-BJT is defined in the context of MOS manufacturing using a spacer layer structure and corresponding processing steps that have been customized and optimized over decades, a very thin base region can be manufactured with high precision to provide high gain. In other words, the structure of the L-BJT can be customized with high precision based on the characteristics and manufacturing steps of the MOS transistor, not only providing synergy between the corresponding manufacturing steps, but also leveraging the precision of the MOS manufacturing process to improve the performance of the BJT in the BiCMOS platform.
[0184] Fig. 7A shows a PNP L-BJT 240 (previously in Figure 2201 and 202. In some embodiments, the PNP L-BJT 240 may include an emitter region 206, a base region 201, a collector region 211, and a layer stack formed on a portion of the collector region 211. The emitter region 206 includes a highly doped (HD) semiconductor material having a first polarity or semiconductor type (e.g., p-type or n-type) and is formed in a lightly doped (LD) base well 208 having a second polarity or semiconductor type opposite to the first polarity or semiconductor type. In some cases, the HD emitter region 206 and the LD base well 208 may include the same semiconductor material doped with opposite types (n-type and p-type) of dopants. The collector region 211 may include at least a portion of the drift region 205, the HD collector region 707, and the LD collector region 710. In some examples, the collector region 211, the LD collector region 710 form a common well 212. The HD and LD collector regions 707, 710 and the common well 212 may include semiconductor materials of the same type (p-type or n-type). In some embodiments, the layer stack may include a reduced surface field (RESURF) dielectric layer 717b (e.g., an oxide layer such as silicon dioxide), a RESURF layer 714b disposed on the RESURF dielectric layer 717, and a RESURF contact layer 712b disposed on the RESURF layer 714b. In addition, the PNP L-BJT 240 may include a spacer layer 716b disposed on the sidewalls of the stacked layers. The RESURF layer 714b, the RESURF contact layer 712b, and the base and collector spacer layer 216b may include different types of materials. The layer stack may be disposed on the drift region 205 of the BJT 240.
[0185] As described above, the layer stack can be configured as a RESURF layer stack to keep the amplitude of the lateral electric field component along the drift region 205 constant, thereby increasing the operating voltage of the PNP L-BJT 240. Equivalently, the layer stack can allow the doping concentration of the drift region to be increased without reducing the operating voltage of the PNPL-BJT 240. In some cases, the RESURF layer 714b can be maintained at the RESURF potential by connecting the contact layer 714b to a voltage source that provides the RESURF potential. The RESURF potential can be a potential that is less than the potential of the HD collector region 707 (also referred to as the collector potential). In some cases, the RESURF potential can be close to or substantially equal to the potential of the emitter region 206. In some cases, the RESURF potential is substantially equal to the ground potential. In some cases, the operating voltage of the L-BJT may increase in proportion to the difference between the RESURF potential and the collector potential. Equivalently, for a given operating voltage, the doping concentration of the base region 201 and / or the drift region 205 may be increased in proportion to the difference between the RESURF potential and the collector potential.
[0186] Including a layer stack in the structure of the L-BJT provides at least three advantages over conventional BJTs (e.g., V-BJTs) used in BiCMOS platforms, including: (1) the ability to co-fabricate the emitter region 206 of the PNP L-BJT 240 with the HD source region 720 of the NMOS transistor 280 while forming the base region 201; (2) increasing the operating voltage of the L-BJT or increasing the doping concentration of the base region 201 and / or the drift region 205; and (3) fabricating a high-gain L-BJT with a precisely defined short base length.
[0187] In some examples, the entire structure of the L-BJT can be formed in the common well. In some other examples, such as PNP L-BJT 240, the L-BJT can be partially formed in the common well and a second region (e.g., the second well) that includes a semiconductor having an opposite dopant type and / or an opposite polarity as compared to the common well. In these examples, the common well and the second region can have an interface below the base well 208 such that a first portion of the base well 208 shares an interface with the common well 212 and a second portion of the base well 208 shares an interface with the second well. In some cases, an electrical connection can be provided to the LD base region above the second well. In the example shown, PNP L-BJT 240 is a PNP BJT having a p-type HD emitter region (p-type HD emitter region) formed in the n-type LD base well 208. ++ ) 206, and a p-type collector region including a p-type well formed in the p-type well 204, the p-type well including the HD and LD collector regions 207, 210. As shown, in this example, the base LD region 208 is partially formed in the p-well (common well) and the n-type well 202 that shares an interface with the p-well 204 below the LD base region.
[0188] Schematic dopant concentration profile 241 shows the relative doping level along the lateral direction (eg, y-axis) in PNP L-BJT 240. The dopant concentration (D E ) can be greater than the doping concentration in the base region (D B ). The dopant concentration in the HD collector region (D CHD ) can be larger than the LD collector region (D CLD ) in the drift region. DR ) can be smaller than D B and D CLD In some examples, D B and D CLD can be different or substantially equal. In some examples, D E and D CHD can be different or substantially equal. In some cases, DDR Can be from 5×10 15 cm -3 to 5×10 16 cm -3 . D B From 10 16 cm -3 to 10 18 cm -3 , D E and D CHD Can be greater than 10 18 cm -3 .
[0189] Figure 7B A side cross-sectional view of an NMOS transistor 280 and a corresponding dopant concentration profile are schematically shown. In some embodiments, the NMOS transistor 280 may be a field effect (FET) transistor including a drain region, a source region, and a gate stack formed on a channel region 203 between the drain region and the source region. The source region may include a highly doped (HD) source region 220 formed in a common well 212. Similarly, the drain region of the MOSFET 280 may include a highly doped (HD) drain region 220 formed in a lightly doped drain region (LD) region 209, wherein the HD drain region (or source region) and the LD drain region (or source region) include semiconductors of the same type or polarity (n-type or p-type). The gate stack may include a gate dielectric layer 217a (e.g., an oxide layer), a gate layer 214a disposed on the gate dielectric layer 217a, and a gate contact layer 212a disposed on the gate layer 214a. In addition, the NMOS transistor 280 may include a gate spacer 216a disposed on the sidewalls of the gate stack. The gate layer 214 a , the gate contact layer 212 a , and the gate spacer layer 216 a may each include different types of materials.
[0190] In some cases, the HD and LD regions of MOSFET 280 can include the same semiconductor material (e.g., silicon) doped with the same dopant. The source and drain regions are formed in a well that includes a semiconductor doped with a dopant type opposite to that of the source and drain regions. In the example shown, MOSFET 280 is an n-channel MOSFET having n-type drain and source regions formed in a p-type well, each drain region including an n-type HD region (N ++ ).
[0191] Dopant concentration profile 281 shows the relative doping concentration along the lateral direction (eg, y-axis) in MOSFET 280. The doping concentration (D SHD ) can be greater than the doping concentration of the LD source region (D SLD ). Similarly, the dopant concentration in the HD drain region (DDHD ) can be greater than the doping concentration in the LD drain region (D DLD ). The dopant concentration in the channel region (D CH ) can be smaller than D DLD and D SLD In some examples, D SHD and D DHD can be different or substantially equal. In some examples, D SLD and D DLD can be different or substantially equal. In some cases, D CH Can be from 5×10 15 cm -3 to 5×10 16 cm -3 , D SLD and D DLD From 10 16 cm -3 to 10 18 cm -3 , D SHD and D DHD Can be greater than 10 18 cm -3 .
[0192] In some embodiments, in addition to the layer stack and the gate stack, all regions and wells of the PNP L-BJT 240 and the NMOS transistor 280 and the substrate 290 may include single crystal silicon. The isolation dielectric layer 218, the gate dielectric layer, and the RESURF dielectric layer may include a dielectric layer (e.g., an oxide, such as thermally grown or deposited silicon dioxide, or other suitable dielectric layers, such as a high-K dielectric layer), the gate and RESURF layers may include polysilicon or metal, and the gate and RESURF contact layers may include silicide. In some embodiments, the isolation dielectric layer 218 may include an oxide layer (e.g., a silicon dioxide layer) thermally grown on the surface of the substrate 290. In some other embodiments, the isolation dielectric layer 218 may include a layer or region (e.g., a silicon dioxide layer) manufactured by forming a trench in the substrate 290 and depositing an oxide material in the trench.
[0193] In various embodiments, the doping concentration (concentration) of the highly doped (HD) region may be 10 18 cm -3 Up to 10 20 cm -3 , the lightly doped (LD) region can be 10 16 cm -3 Up to 10 18 cm -3 , very lightly doped (VLD) regions can have less than 10 15 cm-3 of doping concentration.
[0194] In some cases, the depth of LD source region 709 and / or base well 208 along the vertical direction (eg, along the z-axis) can be from 0.2 microns to 1.0 microns.
[0195] In some cases, the width of LD source region 709 and / or base well 208 along the lateral direction (eg, along the y-axis) can be from 0.18 microns to 2.0 microns.
[0196] In some cases, the depth of LD drain region 711 and / or LD collector region 710 along the vertical direction may be from 0.2 micrometers to 1.0 micrometers.
[0197] In some cases, the width of LD drain region 711 and / or LD collector region 710 along a lateral direction (eg, along the y-axis) may be from 0.18 microns to 2.0 microns.
[0198] In some cases, the depth of HD source well 208HD and / or emitter region 206 along the vertical direction (eg, along the z-axis) can be from 0.1 microns to 0.2 microns.
[0199] In some cases, the width of HD source well 208 and / or HD emitter region 206 along the lateral direction (eg, along the y-axis) can be from 0.18 microns to 1.0 microns.
[0200] In some cases, the depth of HD drain region 721 and / or HD collector region 707 along the vertical direction may be from 0.1 micrometers to 0.2 micrometers.
[0201] In some cases, the width of HD drain region 721 and / or HD collector region 707 along a lateral direction (eg, along the y-axis) can be from 0.18 microns to 1.0 microns.
[0202] In some cases, the length of channel region 203 and / or drift region 205 along the lateral direction can be from 0.18 microns to 0.6 microns.
[0203] In some cases, the length of the RESURF layer 714b along the lateral direction (eg, along the y-axis) may be substantially equal to the length of the drift region along the lateral direction.
[0204] In some cases, the thickness of gate layer 714a and / or RESURF layer 714b along the vertical direction (eg, along the z-axis) can be from 0.1 microns to 0.3 microns.
[0205] In some cases, the thickness of gate contact layer 712a and / or RESURF layer 712b along the vertical direction (eg, along the z-axis) can be from 0.02 microns to 0.05 microns.
[0206] In some cases, the gate dielectric layer 717a and / or the RESURF dielectric layer 774b can have a thickness in a vertical direction (eg, along the z-axis) of 0.01 micrometers to 0.02 micrometers.
[0207] In some cases, the width of the bottom of the spacers 216a and 216b (eg, at the interface between the spaces 216a, 216b and the dielectric layers 217a, 217b, respectively) in the lateral direction may be from 0.05 microns to 0.15 microns.
[0208] In some cases, the length of base region 201 extending from drift region 205 to HD emitter region 206 in the lateral direction (eg, along the y-axis) (base length L B ) can be 0.03 microns to 0.05 microns, 0.05 microns to 0.07 microns, 0.07 microns to 0.1 microns, 0.1 microns to 0.13 microns, 0.13 microns to 0.15 microns, 0.15 microns to 0.17 microns, 0.17 microns to 0.2 microns, or any value in a range formed by these values.
[0209] In some embodiments, the length of base region 204 is substantially equal to the width of the bottom of spacer layers 216a and 216b.
[0210] Still refer to Fig. 7A and 7B As described above, the PNP L-BJT 240 and the NMOS transistor 280 may be fabricated on a common substrate by fabricating the regions and structures of the NMOS transistor 280 and the corresponding regions and structures of the PNP L-BJT 240 in the same fabrication step or in the same process recipe at different steps. For example, the gate stack and spacer layers of one or more (MOS transistors) may be fabricated together with the RESURF stack and spacer 216b of the PNP L-BJT 240. Subsequently, the emitter region 206 of the PNPL-BJT 240 and the HD source region of one or more p-channel MOS transistors, which may be identical in structure to the n-channel NMOS transistor 280, may be fabricated together. Advantageously, the latter fabrication step simultaneously forms the base region 201 and the emitter region 206 of the PNP L-BJT 240.
[0211] In some examples, the gate stack of NMOS transistor 280 and the layer stack of PNP L-BJT 240 have at least one common physical dimension due to being manufactured together or using the same manufacturing recipe. For example, gate layer 714a can have at least one common physical dimension with RESURF layer 714b, and / or gate contact layer 712a can have at least one common geometric dimension with RESURF contact layer 712b. Similarly, spacer layer 716b can have at least one common physical dimension with gate spacer layer 216a. In addition, in some examples, the spacer layer can include the same material, the RESURF and gate layer can include the same material, and the RESURF and gate contact layer can include the same material.
[0212] In some cases, LD base well 208 and LD source region 709 may have at least one common dimension, HD emitter region 206 and HD drain region 721 may have at least two common dimensions, and HD source region 720 and HD emitter region 206 may have at least one common dimension. In some examples, the LD region and / or the HD region may include the same material with the same or different polarities.
[0213] In some embodiments, at least the LD collector region of the L-BJT, the LD drain region of the MOSFET, and the LD source region may be formed in a single well (e.g., a common well). In some such cases, the HD and LD drain and source regions of the MOSFET may have polarities and / or include dopant types that are different from the polarities and dopant types of the HD and LD collector regions and the HD emitter region of the L-BJT.
[0214] In some cases, spacer layer 216b and spacer layer 216a may have substantially the same dimensions (e.g., within tolerances of the respective manufacturing processes), RESURF layer 714b and gate layer 714a may have approximately the same dimensions, and / or RESURF contact layer 712b and gate contact layer 712a may have substantially the same dimensions.
[0215] In some cases, LD base well 208 and LD source region 709 may have substantially the same size, and / or HD emitter region 206 and HD drain region 721 may have substantially identical sizes.
[0216] In various embodiments, gate layer 714a, RESURF layer 714b, gate layer contact 712a, and RESURF contact layer 714b may include one or more materials having a conductivity greater than 10Ω / sq, greater than 30Ω / sq, less than 50Ω / sq, or greater than 70Ω / sq. In some cases, gate layer 714a, RESURF layer 714b, gate layer contact 712a, or RESURF contact layer 714b may include doped polysilicon or silicide. The spacer layer may include a dielectric material (e.g., silicon dioxide).
[0217] In various implementations, the NMOS transistor 280 and the PNP L-BJT 240 may extend laterally between two isolating dielectric layers 218 (eg, silicon dioxide layers) grown or disposed on a substrate 290 .
[0218] In some embodiments, corresponding features of MOSFET 280 and PNP L-BJT 240 having common physical dimensions can be co-fabricated in the same fabrication steps or using the same process recipe. Figure 2 As described, the HD emitter region 206, LD base well 208, HD collector region 707 and LD 210 region, spacer layer 216b, oxide layer 217b, RESURF layer 214b, and RESURF contact layer 212b of BJT 240 are at least structurally similar to the HD source region 220, LD source region 209a, drain HD 221 and LD 209b regions, gate spacer layer 216a, oxide layer 217, gate layer 214a, and gate contact layer 212a, respectively. However, unlike the HD 220 and LD source 209a source regions of MOSFET 280, the HD emitter region 206 and LD base well 208 are not of the same type. Despite this minor difference, these common features can be formed in concert using the same processing steps on a single substrate. In some cases, one or more regions or layers of the PNP L-BJT 240 and the NMOS transistor 280 may have common physical dimensions, common doping profiles, or common material layers. For example, the layer stack 202 b (RESURF layer 214 b and RESURF contact 212 b) of the PNP L-BJT 240, the layer stack of the NPN L-BJT 235, the gate stack 202 a (gate layer 214 a and gate contact layer 212 a) of the NMOS transistor 280, and the gate stack of the PMOS transistor 230 may have one or more common physical dimensions, including thickness and / or include layers having common materials. In addition, the HD and LD collector regions of the NPN L-BJT 235 may have corresponding implant diffusion regions that have common implant regions with the HD and LD drain regions 721, 720 and 709, 711 of the NMOS transistor 280.
[0219] In some implementations, D SLD and D B Can be basically equal, D DLD and D CLD Can be basically equal, D SHD and D B Can be substantially equal, D DHD and D CHD Can be substantially equal and / or D CH and D DR In some cases, the D of the PNP L-BJT 240 may be substantially equal. DR can be greater than the D CH .
[0220] Figure 7CA side cross-sectional view of another embodiment of an L-BJT 730 is schematically shown. In some cases, the upper limit of the operating voltage of the L-BJT 730 may be greater than the operating voltage of the LBT 240. Similar to the PNP L-BJT 240, the L-BJT 730 includes a laterally arranged emitter-base and base-collector junctions, and may operate primarily based on carrier transport in a lateral direction parallel to the main surface of the substrate on which the L-BJT 730 is manufactured. However, the L-BJT 730 may not have an LD collector region, and may include a thick RESURF dielectric layer 734b in addition to the RESURF dielectric layer 717b. The L-BJT 730 may include one or more features described above with respect to the PNP L-BJT 240. The L-BJT 730 may include a HD emitter region 206 formed in the LD base region. The collector region 736 of the L-BJT 730 is part of the common well 212, which supports bipolar carrier transport to and from the base region 201, and can include the HD collector region 707. In some cases, the length of the drift region 705 of the L-BJT can extend from the HD collector region 707 to the base region 201 in the lateral direction (y-axis). The RESURF dielectric layer of the L-BJT 730 can include a thin RESURF dielectric layer 734a and a thick RESURF dielectric layer 734b. Therefore, a first portion of the RESURF layer 732a of the L-BJT 730 can be disposed on the thin RESURF dielectric layer 734a, and a second portion of the RESURF 732a can be disposed on the thick RESURF dielectric layer 734b. In some cases, the RESURF layer of the L-BJT 730 can be longer than the RESURF layer of the L-BJT 240. Spacer layer 216 may be disposed on the sidewalls of RESURF layer 732a and on thin dielectric layer 734a. Similar to the formation of base region 201 of PNP L-BJT 240, base region 201 of L-BJT 730 may be formed simultaneously with the formation of HD emitter region 206 and using spacer layer 216b. In some cases, the length of the drift region of L-BJT 730 may be substantially equal to the length of PNPL-BJT 240. In some cases, the length of the drift region of L-BJT 730 may be greater than the length of the drift region of PNP L-BJT 240. In some cases, the longer length of drift region 705 and / or the presence of thick RESURF dielectric layer 734b may increase the upper limit of the operating voltage of L-BJT 730 compared to the operating voltage of PNP L-BJT 240. In some cases, extending the length of drift region 705 can significantly increase the upper limit of the operating voltage of L-BJT 730. An example of an L-BJT with a long drift region, also referred to as an L-DBJT with an extended drift region, is discussed below.
[0221] Similar to PNP L-BJT 240, in some embodiments, base well 208 of L-BJT 730 can be partially formed in common well 212 and second region 214. In some cases, the interface between collector region 212 and second region 214 can be aligned with the emitter-base junction of L-BJT 730. In some examples, the entire structure of L-BJT 730 can be formed in common well 212.
[0222] In some cases, L-BJT 730 can be at least partially co-fabricated with a MOS transistor similar to MOS transistor 240. In some examples, L-BJT 730 can be co-fabricated with a MOS transistor having a structure similar to L-BJT 730 (eg, LDMOS). Fig.7D A side cross-sectional view of an LDMOS transistor is schematically shown. The LDMOS transistor 740 includes an LD source region 709 formed in a common well 212, an HD source region 720 formed in the LD source region 709, and an HD drain region formed in the common well 212. The gate dielectric layer of the L-BJT 730 may include a thin gate layer 734c and a thick gate dielectric layer 734d. Therefore, a first portion of a gate RESURF layer 732b of the MOS transistor 740 may be disposed on the thin dielectric layer 734c, and a second portion of the gate layer 732b may be disposed on the thick gate dielectric layer 734d. The LD base region 208, the thin RESURF dielectric layer 734a, the thick RESURF dielectric layer 734b, the RESURF layer 732a, and the HD collector region 707 of the L-BJT 730 may be fabricated together with the corresponding layers and regions of the MOS transistor 740, and may include substantially the same physical dimensions as these layers and regions.
[0223] The dopant concentration profile 231 shows the relative doping level in the L-BJT 730 along the lateral direction (eg, the y-axis). The dopant concentration (D E ) can be greater than the doping concentration in the base region (D B ). The dopant concentration in the drift region (D DR ) can be smaller than D B and D CHD In some cases, the relative doping levels in the lateral direction in LDMOS 740 can be similar to th3 of L-BJT 730. In some cases, the D DR can be greater than the D CH .
[0224] The base well 208 of the PNP L-BJT 240 or L-BJT 730 may include one or more HD base regions having the same polarity or semiconductor type as the base well 208. The HD base regions may be used to provide electrical connections to the base well 208 and the base region 201. In some embodiments, the base regions may be aligned with the HD emitter regions along a lateral direction (e.g., along the x-axis) that is perpendicular to the lateral direction (e.g., the y-axis) in which the emitter-base and base-collector junctions are aligned. In some other embodiments, the base regions may be aligned with the HD emitter regions along a lateral direction (e.g., along the y-axis) that is parallel to the lateral direction in which the emitter-base and base-collector junctions are aligned.
[0225] In some cases, forming base region 208 partially in common well 212 and a second region (e.g., second well) having opposite polarity and aligning the interface between collector region 212 and second region 214 with the emitter-base junction of L-BJT 730 or L-BJT-240 can improve the performance of the BJT.
[0226] Reference Figure 2 and 7A -7D, in some embodiments, the length of the layer stack 202b along the lateral direction (e.g., along the y-axis) may be longer than the length of the gate stack 202a along the lateral direction. In some embodiments, the length of the HD collector region 707 along the lateral direction (e.g., along the y-axis) may be less than the length of the HD drain region of the MOS transistor 230 along the lateral direction.
[0227] Fig. 8A A three-dimensional (3D) cross-sectional view of an example L-BJT 800 having a HD base region 802a disposed in a base well 208 in a corner region between a HD emitter region 206 and a spacer layer 216b disposed on a sidewall of a layer stack 714b is shown. Figure 8B A top two-dimensional (2D) cross-sectional view of the L-BJT 800 is shown in a plane 810 that is parallel to and above the top surface of the substrate 290. The base region 802a and another base region 802b are disposed in two corner regions between the HD emitter region 206 and the spacer layer 216b. The base region 802a, the HD emitter region 206, and the base region 802b are disposed near the RESURF layer 714b and are arranged in a lateral direction along the x-axis.
[0228] Figure 8CFIG. 2 shows a 2D cross-sectional view of another example L-BJT 804 having an HD base region 806 disposed in the base well 208 near an edge of the HD emitter region 206, which is opposite to another edge disposed near the spacer layer 216b on the sidewalls of the RESURF layer 714b and the RESURF dielectric layer 717b. The HD base region 806 and the HD emitter region 206 are separated by a gap having a width of 0.1 to 1 micrometer in the lateral direction. The HD base region 806 has the same polarity as the base well 208, but has a higher doping concentration. In some examples, the doping concentration of the HD base region 806 can be greater than 10 20 cm -3 . Fig.8D A top two-dimensional (2D) view of the L-BJT 804 is shown. The base region 806, HD emitter region 206 are aligned in a lateral direction along the y-axis.
[0229] In some cases, the base well 208 of the L-BJT 804 may be partially disposed in the common well 212 and the second region (e.g., second well) 214 having a polarity opposite to the common well 212. The HD emitter region 206, the HD collector region 707, and the common well 212 (e.g., a well shared with a MOS transistor) have the same polarity opposite to that of the base well 208 and the HD base region 806.
[0230] Although PNP L-BJT 240 or L-BJT 804 may be co-fabricated with MOS devices on a common BiCMOS platform, in some embodiments, at least one feature, structure, and doping profile of a co-fabricated L-BJT (e.g., PNP L-BJT 240 and / or L-BJT 804) may be purposefully adjusted, designed, or tailored with corresponding features, structures, and doping profiles of one or more MOS devices (e.g., MOSFET 280 and / or MOSFET 230) to improve the performance of the L-BJT. In some such embodiments, such design, adjustment, or tailoring may result in suboptimal performance of the co-fabricated MOS device. However, the suboptimal performance may be within an acceptable range for an IC device and a corresponding application. In some examples, purposefully adjusting the features, structures, or doping profiles of the L-BJT may take into account its impact on the performance of the co-fabricated MOS device and keep it within an acceptable range for the IC device and a corresponding application.
[0231] In some embodiments, changes made to the fabrication steps of the L-BJT relative to corresponding fabrication steps in the MOS transistor may result in a significant change in the performance of the co-fabricated MOS transistor, e.g., reducing its performance below an acceptable or threshold level. In some such implementations, at least one fabrication step of the L-BJT may be a dedicated fabrication step that is unrelated to the fabrication of corresponding features in the MOS transistor on the same substrate. Dedicated fabrication steps may include, but are not limited to, forming or doping regions (e.g., diffusing dopants), etching structures (e.g., spacer layers), setting layers, etc.
[0232] In some cases, the doping step of the base well 208 of the PNP L-BJT 240 (PNP BJT) can be a dedicated manufacturing step. In some such cases, the doping profile of the base well 208 can be different from the doping profiles of the corresponding drain region 209 and source region 207 of the MOSFET 280. In some examples, the location of the peak doping concentration of the doping profile of the base well 208 along the vertical direction (e.g., the z-axis) can be at or near the lateral emitter-base junction.
[0233] In some cases, spacer layer 216 b of PNP L-BJT 240 may be fabricated (eg, etched) using a dedicated fabrication process that is independent of the process used to fabricate spacer layer 216 of NMOS transistor 280 . Fig. 9A A cross-sectional view of an L-BJT 900 having a spacer 901 fabricated using a dedicated process is shown. For example, the top width, base width, or average width of the spacer 901 of the L-BJT 900 can be smaller than those of a MOS transistor fabricated in the same common well 212 and / or the same substrate. The base width of the spacer 909 can be the length of the base of the spacer 901. The base of the spacer 901 can include a portion of the spacer 901 at the interface between the spacer 901 and the dielectric layer 717b. Advantageously, fabricating the spacer 901 using a dedicated fabrication process can allow for the formation of a base length (L B ) A thin base region less than 0.01 μm, less than 0.05 μm, greater than 0.1 μm or less than 0.15 μm.
[0234] In some cases, emitter region 206 of an L-BJT (eg, PNP L-BJT 240 ) may be fabricated using a dedicated fabrication process that is independent of the process used to fabricate the HD drain or source regions of a MOS transistor on the same substrate (eg, MOSFET 230 ). Fig. 9BA cross-sectional view of an L-BJT 902 having an elongated emitter region 907 fabricated using a dedicated process is shown. For example, the length of the emitter region 907 in a lateral direction (e.g., along the y-axis) can be greater than the length of the HD drain region and the HD source region of a MOS transistor fabricated on the same substrate. Advantageously, the L-BJT having the elongated emitter region 907 can have a lower base current and a higher gain than an L-BJT having an emitter length limited by the length of the HD drain and HD source regions of the co-fabricated MOS device. In some examples, the length of the base well 912 of the L-BJT 902 can be longer than the LD drain and LD source regions of a MOS transistor fabricated on the same substrate to accommodate the elongated emitter. Therefore, both the elongated emitter region 907 and the elongated base well 912 of the L-BJT 902 can be fabricated using a dedicated fabrication process or steps. In some examples, the length of the elongated emitter region can be from 0.18 microns to 0.6 microns. In some examples, the length of the elongated base pores may be 0.18 to 1.0 microns.
[0235] In some cases, the emitter region of an L-BJT (e.g., PNP L-BJT 240) can be extended in a vertical direction (e.g., along the z-axis) to increase the gain of the L-BJT. In this case, a vertical emitter portion can be fabricated on the HD emitter region of the L-BJT using dedicated fabrication steps. Fig. 9C A cross-sectional view of an L-BJT 904 having a vertical emitter portion 903 is shown. In some cases, the vertical emitter portion 903 can be manufactured using a dedicated manufacturing process. The vertical emitter portion 903 can include polysilicon. In some cases, the vertical emitter portion 903 can be disposed on an elongated emitter region 907 of the L-BJT 902. Fig.9D A cross-sectional view of an L-BJT 906 having an elongated emitter region 907 , an elongated base well 912 , and a vertical emitter region 903 is shown.
[0236] In some embodiments, a dielectric layer 905 can be formed between the emitter region 206 and the vertical emitter region 903. The dielectric layer 905 (also referred to as an interface dielectric layer) can be formed naturally during the manufacture of the vertical emitter portion 903, manufactured using a separate process, or a combination of the two. In either case, the thickness of the dielectric layer (in the vertical direction) can be precisely controlled. In some examples, the thickness of the interface dielectric layer can be from 1 pm to 200 pm. In some cases, the thickness of the dielectric layer 905 can be less than a threshold limit to allow carriers to be transported between the emitter region 206 and the vertical emitter region 903 by quantum tunneling. In some cases, the dielectric layer 905 can include silicon dioxide.
[0237] In various embodiments, a dedicated process may be used to fabricate the vertical emitter portion 903 and the corresponding dielectric layer 905 at the base of the vertical emitter region 903 .
[0238] In some cases, the L-BJT (e.g., PNP L-BJT 240) can be vertically isolated from a lower region of a substrate (e.g., substrate 290) below a base well (e.g., base well 206) and / or an LD collector region (e.g., LD collector region 710) of the L-BJT. In some such cases, a buried dielectric layer (e.g., a buried oxide layer) can vertically isolate the L-BJT from the bulk substrate. Such an L-BJT can be fabricated, for example, in a silicon-on-insulator (SOI) substrate. In some cases, a combination of an isolation layer (e.g., isolation layer 218) and a vertical isolation layer formed on the substrate surface can completely isolate the L-BJT. Fig.9E A cross-sectional view of a fully isolated L-BJT 908 is shown, which is laterally isolated by isolation layer 218 and vertically isolated by buried dielectric layer 909. In some examples, buried dielectric layer 909 can be in contact with dielectric layer 218. In some examples, the vertical gap or distance (e.g., along the z-axis) between the bottom surface of dielectric layer 218 and the top surface of buried dielectric layer 909 can be less than 150 nm, less than 100 nm, or less than 50 nm.
[0239] In some embodiments, the above Figures 9A-9F The base well 208, 912 of the described L-BJT design can be formed partially in the common well 212 and a second region (e.g., a second well) that includes a semiconductor having an opposite dopant type and / or an opposite polarity compared to the common well 212. In these examples, the common well 212 and the second region 214 can have a vertical interface 725 (as shown by the dashed line) below the base well 208, 912. In some cases, the vertical interface 725 can be aligned with the vertical emitter-base junction of the L-BJT.
[0240] In some embodiments, the base region of the L-BJT can be formed and the corresponding base length can be precisely defined by thermal diffusion of dopants from the base well under the RESURF layer. In these embodiments, the base width of the spacer layer used to form the emitter region of the L-BJT is reduced to allow dopants to diffuse into the substrate area under the RESURF layer. Fig.9F 9 shows a cross-sectional view of an L-BJT 910 having a thermally diffused base region 911 formed by thermally diffusing dopants laterally (e.g., along the y-axis) from the base well 208 beneath the RESURF layer 714b. In this case, the base length L is precisely defined by controlling the temperature of the sample after forming the base well 208. BThe base width of spacer layer 901 is reduced to bring emitter region 206 closer to the edge of RESURF layer 714b to allow thermal diffusion of dopants from base well 208 under RESURF layer 714b. In some cases, fabrication of spacer layer 901 and thermal diffusion of dopants can be performed as a dedicated process during fabrication of L-BJT 910 on a BiCMOS platform that includes L-BJT 910 and one or more MOS transistors on a common substrate.
[0241] Figures 10A-10D Schematically illustrates cross-sectional views of intermediate structures at some steps in a L-BJT manufacturing process according to an embodiment. Fig. 10A A cross-sectional view of a partially fabricated L-BJT structure is shown during the formation of LD collector region 710. In some cases, isolation oxide layer 218, RESURF oxide layer 717b, RESURF contact layer 712b, and RESURF layer 714b may have been fabricated using standard CMOS fabrication processes. In some examples, these features may be co-fabricated with corresponding features of one or more MOS transistors.
[0242] Fig. 10A The manufacturing steps shown in can include depositing a patterned layer, such as a photoresist layer, and photolithographically patterning the photoresist layer to produce a patterned photoresist layer 1002 covering portions of the isolation dielectric layer 218 and the RESURF contact layer 712b, leaving the edge of the layer stack near the collector region uncovered. After patterning the photoresist layer, an LD collector region can be formed in the common well 212, for example by implanting n-type dopants (for NPN L-BJT) or p-type (for example, for PNP L-BJT). The dopants selectively penetrate into the areas of the common well 212 covered by the unprotected portions of the RESURF oxide layer 717b and form the LD collector region 710.
[0243] The length or width of the collector region 710 in the lateral direction can be determined by the length or width of the unprotected portion of the RESURF oxide layer 717b. The depth of the collector region 710 in the vertical direction (perpendicular to the main surface of the substrate) can be determined by the energy of the implanted dopant and the thickness of the RESURF oxide layer 717b in the vertical direction.
[0244] Fig. 10BThe manufacturing steps shown may include removing the patterned photoresist 1002, depositing a second patterned layer, such as a second photoresist layer, and photolithographically patterning the second photoresist layer to produce a second patterned photoresist layer 1004 covering the isolation dielectric layer 218 and a portion of the RESURF contact layer 712b, leaving the edge of the layer stack near the emitter / base region uncovered. After patterning the second photoresist layer, an LD base well 208 may be formed in the common well 212, such as by implanting p-type dopants (for NPN L-BJT) or n-type (for example, for PNP L-BJT). The dopants selectively penetrate into the areas of the common well 212 covered by the unprotected portions of the RESURF oxide layer 717b and form the base well 208. The length or width of the base well 208 in the lateral direction may be determined by the length or width of the unprotected portions of the RESURF oxide layer 717b. The depth of the base well 208 in the vertical direction (perpendicular to the main surface of the substrate) may be determined by the energy of the implanted dopants and the thickness of the RESURF oxide layer 717 b in the vertical direction.
[0245] Fig. 10C The fabrication steps shown may include removing the patterned photoresist 1004 and depositing a dielectric layer 1006 covering all areas of the L-BJT. In some cases, the dielectric layer 1006 may be formed using a precursor such as tetraethoxysilane (TEOS) and may be deposited using a suitable process such as chemical vapor deposition (CVD), for example, low pressure chemical vapor deposition (LPCVD). The dielectric layer 1006 may have a suitable thickness of between 0.1-0.3 microns in the vertical direction in the L-BJT region, which is related to the final width of the spacer layer formed thereby, which in turn defines the base width as described above. Depending on the consistency of the process used to form the spacer layer, the thickness of a portion of the dielectric layer 1006 on the sidewalls of the layer stack may be 1.5-3 times greater (depending on the thickness of the layer stack).
[0246] Fig. 10DThe fabrication steps shown may include etching the dielectric layer 1004 to create the spacer layer 216b on the sidewalls of the layer stack. In some cases, the etching process may be an anisotropic or directional etching process (e.g., plasma etching) that etches the dielectric layer 1006 in a vertical direction (e.g., along the z-axis). By carefully adjusting the penetration or etching depth of the etching process, a portion of the dielectric layer 1006 disposed on the isolation dielectric layer 218 and a portion of the underlying dielectric film 1006 and the RESURF dielectric layer 717b may be completely removed, thereby exposing the HD base well 208 and the LD collector region 710. At the same time, given the larger thickness of a portion of the dielectric layer 1006 near and at the sidewalls of the layer stack, the spacer layer 216b is retained and formed around the layer stack. Since the etching rate may decrease in proportion to the depth, the spacer layer 216b may have a tapered shape from the wide base of the spacer layer (i.e., in contact with the RESURF dielectric layer 717b) toward the top of the spacer layer near the RESURF contact layer 712b.
[0247] Fig.10E The final fabrication step shown may include depositing a third patterned layer, such as a third photoresist layer, and photolithographically patterning the third photoresist layer to produce a third patterned photoresist layer 1008 covering portions of the isolation dielectric layer 218. After patterning the third photoresist layer, the HD emitter region 206 and the HD collector region 707 may be fabricated on the base well 208 and the LD collector region 710, respectively. The HD emitter region 206 and the HD collector 707 regions are doped by implanting n-type dopants (for NPN L-BJT) or p-type dopants (for example, for PNP L-BJT). The layer stack of the L-BJT and the spacer layer 216b naturally act as a mask for defining the HD emitter region 206 and the HD collector 707 regions. The length of the HD emitter region 206 and the HD collector region 707 is determined by the lateral distance (e.g., along the y-axis) between one of the isolation dielectric layers 218 and the spacer layer 216b. The length L of the base region 201 B is substantially equal to the base width of the spacer layer 216b above the base region 201. The base region 201 is formed using the layer stack as a mask for fabricating the base well 208, and the layer stack and the spacer layer 216b are used as a mask for fabricating the emitter region 206. Therefore, the base length L B The accuracy is limited by the dimensional accuracy of the spacer layer 216b and the layer stack, and the self-alignment of the base well 208 and emitter region 206 with the spacer layer 216b and the layer stack, both of which have been highly optimized over decades of development in CMOS manufacturing.
[0248] The depth of the HD emitter region 206 and the HD collector region 707 along the vertical direction may be determined by the energy of the implanted dopants. After the dopant implantation, the third patterned photoresist layer 1008 may be removed.
[0249] In some embodiments, Figures 10A-10F In one or more manufacturing steps shown, corresponding regions or structures of the MOS transistor may be fabricated on a common well 212 and / or substrate 290 that together fabricate the L-BJT and MOS transistors.
[0250] In some embodiments, the base well 208 ( Fig. 10B ) can be a dedicated implantation process that is not used for the co-fabrication of MOS transistors. In such an embodiment, the doping profile of the base well 208 can be different from the doping profile of the corresponding region (e.g., drain or source region) of the MOS transistor. For example, when the doping profile includes a Gaussian profile, the peak of the Gaussian profile can be located at the lateral emitter-base interface, while for the corresponding source / drain region, the peak can be lower than the HD / LD interface.
[0251] As described above, in some embodiments, the L-BJT may have an elongated emitter region. In these embodiments, one or more masks used to define the length of the RESURF dielectric layer 717b and / or the position of the layer stack relative to the isolation dielectric layer 212 may include features of different sizes for the L-BJT and MOS transistors. Fig. 10A The doping implantation process of an L-BJT having an elongated emitter region 907 and a base well 912 is shown. As shown, the lateral distance between the spacer layer 216b and the isolating dielectric layer on the emitter side (left) is longer than the lateral distance on the collector side (right). As a result, the base well 912 is longer than the LD collector region (in a lateral direction parallel to the y-axis), and the HD emitter region 907 (elongated emitter region) is longer than the length of the HD collector region 707 and / or the length of the drain / source regions of one or more MOS transistors, or is co-fabricated on a common substrate or common well. In some cases, the length of the base well of the L-BJT can be substantially equal to the length of its LD collector region, but the length of its HD emitter region can be longer than the length of the HD collector region 707. In some cases, the upper limit of the length of the elongated emitter 907 can be determined at least in part based on the recombination rate or diffusion length of the carriers in the emitter region. In some examples, the lower limit of the length of the emitter or elongated emitter can be determined by the length (along the y-axis) of the conductive contact electrode disposed on the emitter region 907 or 206. For example, the length of the conductive contact electrode can be 1.5 to 2 times smaller than the length of the emitter region 907 or 206.
[0252] In some embodiments, the isolation dielectric layer 218 can be a thermally grown oxide (also known as local oxide of silicon or LCOS). In some other implementations, the isolation dielectric layer surrounding the L-BJT can be fabricated by etching trenches in the substrate 290 and / or the common well 212 and depositing a dielectric material (e.g., silicon dioxide) in the trenches (also known as shallow trench isolation or STI). Figures 11A-11F Schematically illustrates selected manufacturing steps of an L-BJT having an isolation dielectric (STI) layer 1100 disposed in a trench. In some cases, the RESURF dielectric layer 717b can be fabricated as part of a dielectric deposition process used to fabricate the isolation dielectric layer 1100. In some cases, Figures 11A-11F The manufacturing steps shown are respectively Figures 10A-10F The manufacturing steps shown are similar. For the sake of brevity, the details of the similarities are not reproduced here.
[0253] In some cases, after removing the last patterned photoresist layer, Fig.10E , 10F A capping dielectric layer 224 is disposed on the final structure shown in FIGS. 11A to 11E and 11F. In some cases, the capping dielectric layer 224 may include silicon dioxide.
[0254] In addition, refer to Figures 10A-10E The fabrication process described may include fabricating a HD base region above the base well 208 to provide an ohmic contact with the base region 201. Figures 8A-8C As described above, the HD base region may be fabricated in a corner region between the emitter region 206 and the spacer layer 216 b, or in the vicinity of an edge of the emitter region 206 opposite to another edge near the spacer layer 216 b. Figure 8C As shown, in the latter case, base well 208 may be partially formed in common well 212 and second region 208 having opposite polarity relative to common well 212. In some cases, HD base regions 802a / b or 806 may be fabricated using a dedicated fabrication process during which features of MOS transistors are not fabricated.
[0255] In some cases, the electrical connections to the emitter, base, and collector regions can include ohmic contacts to conductive lines (eg, electrodes). In some cases, at least one conductive line can include polysilicon of a doping type similar to that of the region where polysilicon is disposed on the L-BJT.
[0256] As described above, in some embodiments, the emitter region of the L-BJT may be further extended by adding a vertical emitter region to the emitter region, which may further improve the gain of the L-BJT. Figures 12A-12C The manufacturing process of the vertical emitter portion 908 on the emitter region 206 of the L-BJT is schematically shown. Fig. 12A A cross-sectional view of a partially fabricated L-BJT structure is shown during formation prior to fabrication of the vertical emitter portion 908. The emitter region 206, base well 208, HD and LD collector regions 707, 710, isolation oxide layer 218, RESURF oxide layer 717b, RESURF contact layer 712b, RESURF / layer 714b, and capping dielectric layer 224 may be fabricated using standard CMOS fabrication processes. In some examples, these features may be co-fabricated with corresponding features of one or more MOS transistors. Fig. 12B The fabrication steps shown may include depositing a photoresist layer and photolithographically patterning the photoresist layer to cover the capping dielectric layer 224 except for the area above the emitter region 206 through which a hole is etched to expose the corresponding area of the emitter region 206. Fig. 12C In the next manufacturing step shown, a vertical emitter region 908 is fabricated by depositing a semiconductor material having a medium to high conductivity. In some cases, the vertical emitter portion may include polysilicon. In some cases, the vertical emitter portion may include polysilicon having a dopant of the same type (p or n) as the dopant used to dope the emitter region 206. In various examples, the vertical emitter portion 908 may be fabricated prior to a metallization process that may electrically connect the base, emitter, or collector of the eth L-BJT to more components or devices. In some cases, the vertical emitter portion 908 may serve as an ohmic contact between the metallization layer and the emitter region 206.
[0257] In some examples, a thin interface dielectric layer 905 (e.g., an oxide layer) can be grown on the base and emitter regions 206 of the vertical emitter portion 908. The thin interface dielectric layer 908 can increase the gain of the L-BJT by reducing the carrier gradient in the emitter region 206 and reducing the base current. In some cases, the interface dielectric layer 908 can be a growth layer formed using a controlled growth process (such as oxidation or deposition), can be a native oxide, or a combination thereof. In some cases, the thickness of the interface dielectric layer 908 can be as small as 2 atoms of oxide material. In some examples, the thickness of the interface dielectric layer 908 can be determined based on the desired gain of the corresponding L-BJT. In some other examples, the thickness of the interface dielectric layer can be optimized to provide a maximum level of gain within the constraints imposed by other parameters and features of the L-BJT.
[0258] Figures 12D-12F Schematically illustrates selected fabrication steps of an L-BJT including a vertical emitter portion and having an isolation dielectric layer 1100 disposed in a trench (rather than a thermally grown field oxide). In some cases, Figures 12D-12F The manufacturing steps shown are respectively Figures 12A-12C The production steps shown are the same.
[0259] As described above, in some embodiments, the base region of the BJT can be a thermally defined base region (rather than defined by the base width of the spacer layer). FIGS. 13-13C schematically illustrate the fabrication process of the thermally defined base region 901 and the corresponding base well 918 and HD collector region 919. Fig.13A A cross-sectional view of a partially fabricated L-BJT structure is shown during formation of the HD collector region 919 and prior to fabrication of the base well 918. In some cases, the isolation oxide layer 218, the RESURF oxide layer 717b, the RESURF contact layer 712b, and the RESURF layer 714b may have been fabricated using standard CMOS fabrication processes. In some examples, these features may be co-fabricated with corresponding features of one or more MOS transistors.
[0260] Fig.13A The manufacturing steps shown in can include depositing a photoresist layer and photolithographically patterning the photoresist layer to produce a patterned photoresist layer 1002 covering the isolation dielectric layer 218 and a portion of the RESURF contact layer 712b, leaving the edge of the layer stack 202b near the collector region uncovered. After patterning the photoresist layer, an initial LD collector region 1310 can be formed in the common well 212, for example by implanting n-type dopants (for NPN L-BJT) or p-type (for example, for PNP L-BJT). The dopants selectively penetrate into the areas of the common well 212 covered by the unprotected portions of the RESURF oxide layer 717b and form the initial LD collector region 1310.
[0261] Fig. 13B The fabrication steps shown may include removing the patterned photoresist 1002, depositing a second photoresist layer, and photolithographically patterning the second photoresist layer to produce a second patterned photoresist layer 1004 covering portions of the isolation dielectric layer 218 and the RESURF contact layer 712b, leaving the edges of the layer stack 202b near the emitter / base region uncovered. After patterning the second photoresist layer, an initial LD base well 1308 may be formed in the common well 212, for example by implanting p-type dopants (for an NPN L-BJT) or n-type (e.g., for a PNP L-BJT).
[0262] The depth of the initial base well 1308 and the initial collector well 1310 in the lateral direction (eg, along the y-axis) may be based on the desired depth of the base well 918 and the length (L) of the thermally diffused base region 901. BThe length of the initial base well 1308 and the initial collector well 1310 along the vertical direction (eg, along the z-axis) may be based on the desired length of the base well 918 and the length of the heat diffusion base region 901 (L B ) to determine.
[0263] Fig. 13C The fabrication steps shown include removing the patterned photoresist 1004 and increasing the temperature of the structure to form the base well 918, the thermally diffused base region 901, and the HD collector region 919 by thermally diffusing dopants from the initial base well 1308 and the initial collector well 1310. As a result of the lateral diffusion of dopants from the initial base well 1308, the base length (L B ) and the base depth are determined by the doping profile of the initial base well 1304 and the temperature increase process. In some cases, the temperature increase process for diffusing the initial base well 1308 and the LD collector region 1310 can include increasing the temperature of the structure from an initial temperature to a final temperature. In some examples, the temperature can be continuously increased in a linear manner, a nonlinear manner, or a combination of both. In some examples, the temperature can be gradually increased if the temperature is kept constant or nearly constant in each step to allow the dopant to diffuse before the next temperature increase. In some cases, the length of the heat-diffused base region 901 in the lateral direction (e.g., the lateral diffusion length of the dopant below the layer stack 202b) can be 50nm to 100nm, 100nm to 150nm, 150nm to 200nm, or any range formed by these ranges, or a larger or smaller range.
[0264] In some cases, the doping profile of the initial base well 1308 can be designed so that when the thermal diffusion process is completed, the doping profile of the resulting base well 918 matches or is substantially the same as the target doping profile in the lateral and vertical directions. In these cases, the thermal diffusion process can simultaneously define L B and the location of the peak dopant concentration along the vertical axis.
[0265] After fabricating the base well 918 and the LD collector region 919, the above-described Fig. 10C The fabrication steps depicted deposit a dielectric layer 1006 (eg, a layer comprising TEOS) over the L-BJT structure. Fig.13DThe fabrication steps shown may include a dedicated etching step for etching dielectric layer 1006 and a portion of RESURF dielectric layer 717b over base well 918 and LD collector region 919. In some cases, a very thin spacer layer 901 may be left after an etching process (e.g., a vertical etching process). In some cases, the base width of spacer layer 901 may be 1 nm to 5 nm, 5 nm to 10 nm, 10 nm to 20 nm, 20 nm to 30 nm, 30 nm to 40 nm, or 40 nm to 50 nm. Fig.13E The final manufacturing steps shown can be similar to those described above with respect to Fig.10E The final manufacturing step described, in which the HD emitter region 206 and the HD collector region 707 are manufactured by implanting n-type dopants (for NPN L-BJT) or p-type dopants (for example, for PNPL-BJT). The layer stack 202b of the L-BJT and the thin spacer layer 901 are used as masks to define the HD emitter region 206 and the HD collector region 707. Although it is possible to manufacture an L-BJT with a heat-diffused base region 911 without the need to manufacture a spacer layer on the sidewalls of the layer stack 202b, in some cases, in order to maintain the synergy between the manufacturing processes of the L-BJT and the MOS transistor and reduce the manufacturing steps, the dielectric layer 1004 can still be set on the L-BJT structure (after the base region is heat-diffused). In this case, the L-BJT is co-manufactured (at least partially) with the MOS transistor on a common substrate. However, in some cases (for example, when the L-BJT is not co-manufactured with the MOS transistor), the deposition of the dielectric layer 1004 after the heat-diffused base region 911 can be skipped. In this case, the etching process ( Fig.13D ) removes the portion of the RESURF dielectric 717b layer disposed on the base well 918 and the LD collector region 919, but does not form a spacer structure. Fig.13E ), the layer stack 202b acts as a mask for defining the HD emitter region 206 and the HD collector region 707.
[0266] The edge of the layer stack 202b is self-aligned with the edge of the initial base well 1308, and the edge of the spacer layer 901 is self-aligned with the emitter region 206. As a result, the emitter and base junctions are self-luminous relative to the edge of the spacer layer 901 or the edge of the layer stack 202b (in the absence of the spacer layer). The base length and the position of the base-collector junction are determined by the thermal diffusion process. Therefore, by controlling the doping profile of the initial base well 1308 and the temperature rise process, the doping profile and the length of the thermally diffused base region 901 can be precisely adjusted.
[0267] The doping profile of the thermally diffused base region 901 may include a spatially varying dopant concentration that decreases (linearly or nonlinearly) from the edge of the layer stack 202b laterally toward the LD collector region 919. In contrast, a base region defined by a spacer layer (e.g., Figures 10A-10E The doping profile of the base region 201 in the process shown may include an almost constant or slowly varying dopant concentration (compared to the dopant density variation in the thermally diffused base region 901) in the lateral direction from the edge of the layer stack 202b toward the LD collector region 919. Advantageously, the doping profile of the thermally diffused base region 901 increases the operating speed of the corresponding L-BJT compared to an L-BJT having a base region with an almost constant or slowly varying dopant concentration.
[0268] Figures 14A-14D Schematically illustrates selected fabrication steps of an L-BJT having an isolation dielectric layer 1100 disposed in a trench. In some cases, the RESURF dielectric layer 717b can be fabricated as part of a dielectric deposition process used to fabricate the isolation dielectric layer 1100. In some cases, Figures 14A-14E The manufacturing steps shown are similar to Figures 13A-13E Fabrication steps shown. Lateral bipolar junction transistor with extended collector drift region
[0269] Fig.15A Schematically shows Fig.9D906, and the variation of the voltage 1502 and the electric field 1506 along the drift region and the base region 201, 205 of the L-BJT 906. The drift region 205 of the L-BJT 906 extends laterally from the vertical boundary of the base well 208 to the vertical boundary of the LD collector region 710, which is aligned with the first edge of the layer stack 202b, and the vertical edge is aligned with the second edge of the layer stack 202b opposite to the first edge. The layer stack 202b includes a RESURF connector layer 712b, a RESURF layer 714b, and a RESURF dielectric layer 734a. The RESURF connector layer 712b is disposed on and electrically contacts the RESURF layer 714b, and the RESURF layer 714b is disposed on the RESURF dielectric layer 734a that electrically isolates the RESURF layer 714b from the drift region 205. The RESURF layer 714b can modify the electric field (E-field) in the drift region 205 by supporting a vertical electric field component, e.g., maintaining a substantially constant or nearly constant amplitude of a lateral electric field component along the drift region 205. In some cases, e.g., when the RESURF connector layer 712b is electrically biased to a voltage having an amplitude lower than a voltage provided to the HD collector region 707, the amplitude of a lateral component of the E-field 1506 generated by applying a voltage between the HD collector region 707 and the HD emitter region 206 (also referred to as an emitter region) can be maintained substantially constant or nearly constant along the drift region 205 and the base region 201. In some examples, the constant amplitude of the E-field 1506 (e.g., the lateral component of the E-field) can be substantially equal to a critical E-field amplitude (E critical ), which may result in one or both of reverse biased junction breakdown in the emitter-base junction and dielectric breakdown of the RESURF dielectric layer 734a.
[0270] For example, when the potential difference between the HD collector region 707 and the HD emitter region 206 is approximately 5 volts, the voltage 1502 may decrease monotonically, e.g., substantially linearly, from 5 volts at the HD collector region 707 to nearly zero at the base emitter junction. Thus, the L-BJT 906 may control (e.g., switch) a signal including a 5 volt voltage. In some cases, the omission of the RESURF layer 714b and the RESURF contact 712b in the L-BJT 906 may cause the magnitude of the electric field 1504 to increase monotonically, e.g., substantially linearly, along the base region 201 and the drift region 205. As a result, in the absence of the RESURF layer 714b and the RESURF contact 7121b, a voltage difference of approximately 5 volts between the HD collector region 707 and the HD emitter region 206 may cause one or both of the reverse biased collector-base junction and the collector-base junction RESURF dielectric layer 734a to break down.
[0271] In some cases, refer to Fig.9E , 9F The role of the RESURF layer of L-BJTs 908, 910, 902, 904, 240 described in , 9B, 9C and 7A can be similar or the same as the role of the RESURF layer 714b of L-BJT 906, for example, to keep the amplitude of the lateral electric field component constant or nearly constant along the corresponding drift region. Similarly, the changes in the electric field and the corresponding voltage along the drift region of L-BJTs 908, 910, 902, 904, 240 and 235 can be similar to Fig.15A A variation of the L-BJT 906 shown in FIG.
[0272] Fig. 15B Schematically shows Figure 7C , and the variation of voltage 1506 and electric field 1506 along the drift region and base region 201, 705 of the L-BJT 730. The drift region 705 of the L-BJT 730 extends laterally from the vertical boundary of the base well 208 aligned with the edge of the RESURF layer 732a to the vertical boundary of the HD collector region 707 aligned with the end of the thick isolation dielectric layer 734b. When the thick isolation dielectric layer 734b is a LOCOS layer, the end may correspond to its "bird's mouth". The RESURF layer 732a includes a first portion 1510a on the RESURF dielectric layer 734a and a second portion disposed on the thick isolation dielectric layer 744b. The RESURF dielectric layer 734a and the thick isolation dielectric layer 734b electrically isolate the RESURF layer 732a from the drift region 705. In some cases, the first portion 1510a and the second portion 1510b of the RESURF layer 732a are referred to as the first and second RESURF plates or field plates of the L-BJT 730. The RESURF layer 732a (first and second field plates) can change the electric field (electric field) in the drift region 705 by supporting the vertical electric field component, for example, keeping the amplitude of the lateral electric field component substantially constant or nearly constant along the drift region 705. The RESURF layer 732a is generally formed of a conductive material and may have a conductivity greater than 10Ω / sq, greater than 30Ω / sq, less than 50Ω / sq, or greater than 70Ω / sq. For example, the RESURF layer 732a may include polysilicon (e.g., n-doped or p-doped polysilicon). However, the embodiment is not limited thereto, and the RESURF layer 732a may be formed of a metal.
[0273] In some cases, a RESURF connector layer (not shown) can be disposed on the RESURF layer 732a to provide electrical contact with the RESURF layer 732a. In some cases, for example, when the RESURF layer 732a is electrically biased (e.g., via the RESURF contact layer) to a voltage having an amplitude less than the amplitude of the voltage provided to the HD collector region 707, the amplitude of the lateral component of the electric field 1506 generated in the drift region 705 and the base region 201 (caused by the voltage applied between the HD collector region 707 and the HD emitter region 206) can remain substantially constant along the drift region 705 and the base region 201. Thus, the corresponding voltage can decrease monotonically (e.g., substantially linearly) along the drift region 705 and the base region 201 toward the emitter-base junction. In some examples, the constant amplitude of the lateral component of the electric field 1506 can be substantially equal to the critical electric field amplitude (E ) corresponding to reverse junction breakdown. critical The length of the drift region 705 of the L-BJT 730 in the lateral direction may be greater than the length of the drift region 205 of the L-BJT 240 ( Fig.27A ). Therefore, the drift region 215 and the RESURF layer (field plate) 714b ( Fig.15A ), the drift region 705 and the field plates 1510 a and 1510 b of the L-BJT 730 can support a larger voltage drop from the HD collector region 707 to the HD emitter region 206 .
[0274] As an example, when the potential difference between the HD collector region 707 and the HD emitter region 206 is approximately 20 volts, the voltage 1502 can decrease monotonically (e.g., substantially linearly) from 20 volts at the HD collector region 507 to nearly zero at the base-emitter junction. Thus, the L-BJT 730 can control (e.g., switch) a signal including a voltage of 20 volts. In some examples, by extending the drift region 705 and one or both of the field electrodes 1510a, 1510b, the voltage drop along the drift region 705 and the base region 201 can be increased to 80 volts, thereby allowing the L-BJT 730 to control (e.g., switch) a signal including a voltage with a magnitude of 80 volts.
[0275] refer to Fig. 15B , in some cases, omitting the RESURF layer 732a in the L-BJT 730 may cause the magnitude of the electric field 1504 to increase monotonically, e.g., substantially linearly, along the drift region 705 and the base region 201 of the L-BJT 730. As a result, in the absence of the RESURF layer 732a, a voltage difference of 20 volts between the HD collector region 707 and the HD emitter region 206 may cause the emitter-base junction of the L-BJT 730 to break down.
[0276] Still refer to Fig.15A and 15B , in some cases, the L-BJT 906 (which can operate between 0 and 5 volts Fig.15A ) may be referred to as a low voltage (LV) L-BJT, and L-BJT 730, which may operate between zero and 80 volts (e.g., depending on the length of its drift region), may be referred to herein as a medium voltage (MV) L-BJT. A comparison between the voltage drops across the drift regions of L-BJT 730 and L-BJT 906 shows that extending the length of the drift region in the lateral direction and providing a field plate above the extended drift region may increase the magnitude of the voltage drops across the drift region and the base region. Thus, the layer stack and corresponding drift region of the L-BJT not only self-align the emitter region, base well, and collector region of the L-BJT, but also define a drift region of the L-BJT beneath the conductive layer (RESURF layer) that may be tailored to increase the upper limit of the L-BJT operating voltage. Advantageously, both of the above L-BJT designs may be co-fabricated with MOS devices having common structural features using the CMOS co-fabrication techniques described herein. For example, L-BJT 906 can be connected to, for example, low voltage MOS transistor 240 ( Figure 2 ), the L-BJT 730 can be co-fabricated with, for example, a high voltage MOS transistor 740. In some examples, the base well 208 and the HD collector region 707 of the L-BJT 730 and the LD source region 709 and the HD drain region 721 of the MOS transistor 740 can have at least one common physical dimension and / or a common doping profile, and can be co-fabricated on a common substrate or a common well in a common substrate. In some examples, the spacer layer 216b, the RESURF layer 732a, the RESURF dielectric layer 734a, and the thick isolation dielectric layer 734b of the L-BJT 730 and the spacer layer 216a, the gate layer 732b, the gate 734c, and the thick drain dielectric layer 734d of the MOS transistor 740 can have at least one common physical dimension and can be co-fabricated on a common substrate.
[0277] In some examples, the operating voltage (V m ) may be an upper limit of a voltage above which at least one junction and / or at least one dielectric layer within the L-BJT is electrically broken down by an electric field generated across the dielectric layer or junction (e.g., a reverse biased PN junction).
[0278] In some embodiments, the collector region of the L-BJT can be tailored to increase its V m, without substantially changing the size and / or configuration of its emitter and base regions. For example, the collector region of the L-BJT 730 can be further extended by increasing the length of the drift region 705 in the lateral direction (e.g., y-direction), extending the thick isolation dielectric layer 734b, and adding an additional field plate above the extended portion of the thick isolation dielectric layer 734b.
[0279] Advantageously, the fabrication of L-BJTs operating at higher voltages can benefit from modifications to the base region (L B ) lengths, and the co-fabrication of the spacer layers, RESURF dielectric layers, and emitter, base, and HD collector regions with the corresponding regions and layers of the MOS transistor on a common substrate (as described above with respect to L-BJT 240 ( Figure 2 ) as described above). Figure 3B An example of a high voltage (HV) L-BJT 350 with an extended drift region is described. More details about the L-BJT 350 and related L-BJT designs capable of operating at higher voltages (e.g., greater than 80 volts) are described below. In some cases, an L-BJT having an extended drift region and a field plate (e.g., a metal field plate) in addition to a RESURF layer 732a may be referred to herein as an HV L-BJT. In some examples, the HV L-BJT may have a V greater than 80 volts, greater than 90 volts, greater than 100 volts, greater than 150 volts, greater than 200 volts, greater than 250 volts, greater than 300 volts, or greater than 400 volts. m , or a value within the range defined by any of these values.
[0280] Fig.16AA cross-sectional view of an example L-BJT 350 having an elongated drift region (referred to herein as an extended drift region) is schematically illustrated, the drift region including the drift region 205 and the drift region extension 304b. In some cases, the elongated drift region extending between the HD collector region 707 and the base well 206 can be referred to as the drift region of the L-BJT 350. The L-BJT 350 also includes one or more field plates above the layer stack and the drift region extension. The drift region 205 extends laterally from a first end to a second end, and the drift region extension 304b extends laterally from the second end to a third end of the drift region 205. The extended drift region extends from the first end of the drift region 205 to the third end of the drift region extension 304b. The length of the laterally extended drift region can be 0.5 to 2.0 microns, or 2.0 to 5.0 microns, 5.0 to 15.0 microns, 15.0 to 25.0 microns, or a value within a range defined by any one of these values, or less or greater. The length of the drift region 205 in the lateral direction can be 0.18 to 0.6 microns, or a value within a range defined by any one of these values, or less or greater. The length of the extended drift region 304b can be from 0.5 microns to 25.0 microns, or a value within a range defined by any one of these values, or less or greater.
[0281] Similar to L-BJT 240 , HV L-BJT 350 includes emitter region 206 and base region 201 at one end of the extended drift region, and HD collector region 707 at the other end of the extended drift region.
[0282] Still refer to Fig.16A, the layer stack of the HV L-BJT 350 includes a RESURF dielectric layer 734a on the drift region 205, and a RESURF layer disposed on the RESURF dielectric layer 734a, referred to herein as a first field plate 1510a. In some cases, the RESURF connection layer can be disposed on the first field plate 1510a. In some cases, a thick isolation dielectric layer 1614 can be disposed above the drift region extension 304b. The thick isolation dielectric layer 1614 can extend from the edge of the RESURF dielectric layer 734a toward the HD collector region 707. The thickness of the thick isolation dielectric layer 1614 along the vertical direction (e.g., along the z-axis) can be greater than the thickness of the RESURF dielectric layer 734 along the vertical direction. In various embodiments, the difference between the thickness of the thick isolation dielectric layer 1614 and the RESURF dielectric layer 734 can be from 0.01 microns to 0.015 microns. In some examples, the thickness of thick isolation dielectric layer 1614 can be from 0.3 microns to 0.5 microns, within a range of any of these values or less or greater. In various embodiments, the length of thick dielectric layer 1614 can be greater than 20%, 30%, 40%, 50%, or 60% of the lateral distance between the extended drift region length or base well 208 and HD collector region 707.
[0283] Second field plate 1510b can be disposed over a portion of thick isolation dielectric layer 1614. In some cases, second field plate 1510b can be in contact with a top surface of thick isolation dielectric layer 1614. In some cases, first field plate 1510a and second field plate 1510b can include the same material (e.g., a conductive material such as n-doped polysilicon). In some cases, first field plate 1510a and second field plate 1510b can be two portions of a single RESURF layer disposed (e.g., conformally) on a RESURF dielectric layer and at least a portion of thick isolation dielectric layer 1614.
[0284] The extended drift region may extend laterally from a vertical boundary of the base well 208 beneath the first field plate 734a to a vertical boundary of the HD collector region 707). In some cases, the vertical boundary of the base well 208 may be aligned with an edge of the first field plate 734a.
[0285] In some cases, the emitter region 206 (also referred to as the HD emitter region) of the L-BJT 350 can be aligned with the edge of the spacer layer 216b formed on the sidewall of the first field plate 1510a (similar to the L-BJT 240). Therefore, during the manufacturing process, the length of the base region of the L-BJT 350 can be defined by the width of the bottom of the spacer layer 216b. In some cases, at least the base well 208, emitter region 206, spacer layer 216b, first field plate 1510a, and RESURF dielectric layer 734a of the L-BJT 350 can be aligned with a MOS transistor (e.g., MOS transistor 300 ( Figure 3A ) or MOS transistor 280 ( Figure 2 )) are co-fabricated with corresponding regions and features of the L-BJT 350. In some cases, the L-BJT 350 can be co-fabricated with the MOS transistor 300 (e.g., DMOS) on a common substrate (e.g., in a common well 212 formed in the substrate 290). For example, the gate layer in the gate stack 202a of the MOS transistor 300 can be used as a gate contact (corresponding to the first field plate 1510a), and its gate stack extension 302a can include a first drain plate (corresponding to the second field plate 1510b) disposed on the thick drain dielectric layer 1614 (corresponding to the thick isolation dielectric layer) and a second drain plate (corresponding to the third field plate) extending on the thick channel dielectric layer. In some cases, the first, second, and third field plates 1510a, 1510b, 1612, the thick isolation dielectric layer 1614 of the L-BJT 350, and the corresponding plates and layers of the MOS transistor 300 can have at least one common physical dimension and / or material and can be manufactured together (e.g., on a common well in a common substrate) on a common substrate (e.g., in the same manufacturing steps).
[0286] The dopant concentration profile 1616 shows the variation of the average doping concentration level in the L-BJT 350 along the lateral direction (eg, the y-axis). E ) can be greater than the doping concentration in the base region (D B ). The dopant concentration in the extended drift region (D DR ) can be smaller than D B and D DR . The dopant concentration in the HD collector region (D CHD ) can be 10 20 Up to 10 21 cm -3 The dopant concentration (D E ) can be 10 20 Up to 10 21 cm -3 The dopant concentration in the base well (D B ) can be 1018 Up to 10 19 cm -3 In some cases, the dopant profile of the base region may have a peak near or at a lateral boundary between emitter region 206 and base well 208 .
[0287] In some cases, the dielectric layers and field plates in the extended layer stack of the L-BJT 350 may allow for an increase in D compared to an L-BJT with an extended drift region without a field plate. DR , thereby reducing the resistance of the drift region.
[0288] In some cases, in addition to changing the electric field in the extended drift region, the field plate can also dissipate heat generated by the L-BJT 350, thereby improving the L-BJT 350 by allowing greater current to flow through different regions of the L-BJT 350 before current-limiting thermal effects begin.
[0289] Fig. 16B Schematically shows Fig.16A 3 and the electric field variation along the extended drift region of the L-BJT. As described above, the field plates 1510a, 1510b, 1612 generate a vertical electric field component (E RESURF The total electric field (E) at any point along the extended drift region is tot ) can correspond to the emitter-base junction electric field and E RESURF In some cases, at least E tot (E tot-L )1506 remains constant or nearly constant along the extended drift region (in the lateral direction) from the base-emitter junction to the HD collector region. In some examples, E tot The lateral component along the extended drift region, from the HD collector region 707 to the base well 208, may vary by less than 2%, 4%, 6%, 8%, or 10% of its value at the boundary of the base well 208. In some examples, E tot The lateral component of the drift region along the extension from the HD collector region 707 to the base well 208 may remain below a threshold value (eg, E critical ), which may cause breakdown of an oxide layer (e.g., RESURF dielectric layer 734a) or a junction (e.g., base-collector junction) of the L-BJT (e.g., an L-BJT with an extended drift region).
[0290] The voltage drop from the HD collector region 707 to the HD emitter region 206 is equal to the area under the curve 1506 or |E tot-L |×L drift , where L driftis the length of the laterally extending drift region. In some cases, the V m Can be basically equal to |E critical |×L drift , where E critical is an electric field magnitude that can cause electrical breakdown in the emitter-base junction or RESURF dielectric layer 734a. In other words, the required V m The minimum length of the extended drift region or the given L drift Drifting V m Affected by E critical Constraints, E critical depends on the material and thickness of the RESURF dielectric layer, and the material and doping characteristics of the emitter-base junction. In the absence of field plates 1510a, 1510b and 1612, E tot-L The magnitude of 1504 may decrease monotonically, e.g., substantially linearly, along the extended drift region (in the lateral direction) from the base-emitter junction to the HD collector region. tot-L Keep E near the base-emitter junction critical Below or close to E critical , in the absence of a field plate, the length of the drift region should be longer. In the absence of a field plate, V m It can correspond to or be approximated as |E critical |×(L' drift / 2), or E tot-L 1504 line. Therefore, for a given V m and E critical , the length of the extended drift region (L') of L-BJT 350 without a field plate drift ) can be approximately the length of the drift region (L) under the condition of field plate. drift ) times.
[0291] In some embodiments, V may be further increased by extending the extended drift region and adding an additional (eg, fourth) field plate extending over at least a portion of the extended portion of the extended drift region. m In some cases, the thick isolation dielectric layer 1614 can be elongated to provide isolation between the additional field plate and the elongated portion of the extended drift region. Fig. 16CA side cross-sectional view of an HV L-BJT 1600 having a longer drift region extension 304b and a thick isolation dielectric layer 1614 compared to the HV L-BJT 350 is schematically shown. The HV L-BJT 1600 also includes a fourth field plate 1618 that extends laterally from a distal end of the third field plate 1612 relative to the second field plate 1510b toward the HD collector region 707. In some cases, the third and fourth field plates 1612, 1618 are substantially parallel to a top surface of the substrate or a top surface of the thick isolation dielectric layer 1614. A vertical distance h1 (e.g., along the z-axis) between the third field plate 1612 and the top surface of the thick isolation layer 1614 can be greater than a vertical distance h2 (e.g., along the z-axis) between the fourth field plate 1618 and the top surface of the thick isolation film 1614. In some cases, the third and fourth field plates 1612, 1618 can be fabricated in two different metallization layers (e.g., adjacent metallization layers of an IC), with one metallization layer being above the other. In various embodiments, the difference between h1 and h2 can be from 0.5 microns to 1.0 microns. In some examples, h1 and h2 can be 0.5 to 2.5 microns, or values within any range defined by these values, or smaller or larger.
[0292] In various embodiments, the third and fourth field plates 1612, 1618 may cover at least 20%, 30%, 40%, 50%, or 60% of the lateral length of the thick isolation dielectric layer 1614. In various embodiments, the third field plate 1612 may cover at least 20%, 30%, 40%, 50%, or 60% of the lateral length of the thick isolation dielectric layer 1614. In various embodiments, the third and fourth field plates 1612, 1618 may cover at least 20%, 30%, 40%, 50%, or 60% of the lateral length of the extended drift region. In various embodiments, the third field plate 1612 may cover at least 20%, 30%, 40%, 50%, or 60% of the lateral length of the extended drift region.
[0293] In some cases, the doping profile of L-BJT 1600 can be similar to the doping profile 1616 described with respect to L-BJT 350. The placement of the HD base regions of L-BJT 350 and L-BJT 1600 relative to the corresponding HD emitter regions 206 is similar to the placement of the HD base regions 802a / 802b in L-BJT 800 ( Fig. 8A and 8B ).
[0294] The variation of the voltage 1602 and the electric field 1506 along the extended drift region is shown below the corresponding drift region and the drift region extension 304b. Fig. 16CAs shown, the isolation dielectric layer and the field plate can keep the lateral component of the E field 1506 constant or nearly constant along the drift region. Therefore, the voltage 1602 decreases monotonically, for example, substantially linearly, from the HD collector region 707 to the HD emitter region 206. Figure 3B ) is similar, L-BJT 1600's V m Equal to |E critical |×L drift . Given that the L-BJT 1600 drift Longer drift, L-BJT 1600 can have larger V compared to L-BJT 350 m In the example shown, the V of the L-BJT 1600 m About 250V.
[0295] Fig.17 Schematic showing the V of the four different L-BJT designs and structures described in this article m (upper limit of operating voltage) versus drift region length. Low voltage (LV) L-BJT 906 and similar L-BJT designs (e.g., L-BJT 908, 910, 902, 904, 240, or 235) including a single field plate on a thin dielectric layer (RESURF dielectric layer) and / or a drift region length less than 1.0 micron may have a V less than or equal to 5 volts. m Medium voltage (MV) L-BJT 904 and similar L-BJT designs including two field plates, one disposed on a thin dielectric layer and the other disposed on a thick dielectric layer, and / or having a drift region length greater than 1.0 micron but less than 5.0 microns, can have a V greater than 5 volts but less than or equal to 80 volts. m The high voltage (HV) L-BJT 350 and similar L-BJT designs including three field plates, one on a thin dielectric layer and two on a thick dielectric layer and / or having a drift region length greater than 5.0 microns but less than 15.0 microns, can have a V greater than 80 volts but less than or equal to 150 volts. m The HV L-BJT 1600 and similar L-BJT designs that include four field plates, one on a thin dielectric layer and three under a thick dielectric layer, and / or drift region lengths greater than 15.0 microns can have V greater than 150 volts. m .
[0296] In some cases, L-BJTs with drift region lengths less than 1.0 micron can be V m LV L-BJTs less than 5 volts, L-BJTs with drift regions greater than 1.0 micron and less than 5.0 micron can be V mMV L-BJT greater than 5V and less than 80V, L-BJT with offset region length greater than 5.0μm and less than 15.0μm can be V m HV L-BJTs greater than 80 V and less than 150 V, and L-BJTs with drift region length greater than 15.0 μm may be V m Up to 150V or higher HV L-BJTs.
[0297] In some cases, the length of the extended drift region of L-BJT 350 or L-BJT 1600 in the lateral direction can be substantially equal to the length of the first field plate 1510a in the lateral direction. In some cases, the length of the drift region extension 304b of L-BJT 350 or L-BJT 1600 can be substantially equal to the length of the thick isolation dielectric layer 1614.
[0298] In various embodiments, all field plates of the L-BJT are electrically connected to each other. In some examples, the third field plate 1612 can be electrically connected to the second field plate 1510b by a first vertical wire (e.g., a first conductive via), and the fourth field plate 1618 can be electrically connected to the second field plate 1612 by a second vertical wire (e.g., a second conductive via). In some embodiments, the first field plate 1510a and the second field plate 1510b may include polysilicon. In some embodiments, the third and fourth field plates 1612, 1618 may include a metal (e.g., copper, gold, aluminum, chromium, or an alloy made of two or more metals). In various examples, one or more field plates may include a doped semiconductor (e.g., doped polysilicon). In various examples, one or more field plates may include a metal or a metal alloy.
[0299] In some examples, field plates 1510a, 1510b, 1612, and 1618 can be electrically connected to a potential (referred to as a RESURF potential) that has a magnitude less than the magnitude of a potential applied to HD collector region 707. For example, field plates 1510a, 1510b, 1612, and 1618 can be electrically connected to emitter region 206 or a potential provided to emitter region 206. In some examples, field plates 1510a, 1510b, 1612, and 1618 can be electrically connected to a ground potential. In various implementations, the potentials of field plates 1510a, 1510b, 1612, and 1618 can remain constant or nearly constant during operation of the corresponding L-BJT.
[0300] In some embodiments, the extended drift region can be further extended, and more field plates can be added on the extended portion of the extended drift region to further increase the V of the HV L-BJT. mThe vertical distance between the field plate and the top surface of the thick dielectric layer 1614 (or the top surface of the drift region extension 304b) can be less than the vertical spacing between the subsequent field plate and the top surface of the thick dielectric layer 1614. In some cases, the increase in the vertical distance of the subsequent field plate from the surface helps to keep the lateral component of the E field substantially constant along the extended drift region, thereby increasing V m .
[0301] In some examples, field plates 1510a, 1510b, 1612, and 1618, as well as any additional field plates, can be fabricated on different metallization layers (eg, metallization films of an IC).
[0302] In various embodiments, the HV L-BJT with an extended drift region and field plate can be electrically isolated from low voltage or medium voltage electronic devices (e.g., LV and MV L-BJTs and MOS transistors) fabricated on the same substrate to avoid interference or damage to these devices due to accidental coupling of high level voltages from the HV L-BJT. In some cases, the electrical isolation can include completely isolating the HV L-BJT in the lateral and vertical directions to prevent carrier transport or E-field coupling through substantially all areas of the HV L-BJT to reach the substrate area including the low voltage or medium voltage device. In some cases, complete isolation can be achieved using a combination of a buried dielectric layer (e.g., a buried oxide layer) and a deep trench filled with a dielectric material (e.g., a deep trench oxide), or a combination of an isolation dielectric layer near the top surface of the device (e.g., a LOCOS field oxide layer or a shallow trench filled with a dielectric dielectric material).
[0303] Fig.18 A side cross-sectional view of a HV L-BJT 1800 with a fully isolated extended drift region is schematically shown. In some cases, the L-BJT 1800 can be laterally isolated by an isolation dielectric layer 218 (e.g., a field oxide or shallow trench isolation (STI) layer) and vertically isolated by a buried dielectric layer 909 (e.g., a buried oxide layer). In some cases, the L-BJT 1800 can be laterally isolated by a shallow dielectric filled trench (e.g., a shallow trench oxide, also referred to as STI), similar to the above description of Figures 11A-11F The isolation dielectric layer 1100. The L-BJT 1800 may include one or more features described above with respect to the L-BJT 350 or the L-BJT 1600.
[0304] In some examples, the buried dielectric layer 909 can be in contact with the dielectric layer 218. In some examples, the vertical gap or distance between the bottom surface of the dielectric layer 218 and the top surface of the buried dielectric layer 909 (e.g., along the z-axis) can be less than 5 nm, less than 10 nm, or less than 20 nm. In some cases, the extended drift region of the L-BJT 1800 can include the drift region 205 and the drift region extension 304b below the layer stack. The layer stack can include a RESURF dielectric layer 734a and a RESURF layer (or first field plate) 732a disposed on the RESURF dielectric layer 734a. The base well of the L-BJT 1800 can extend vertically from the top surface of the buried dielectric layer 909 to a capping dielectric layer 224, such as a passivation layer, disposed on the L-BJT 1800 along its entire length. The thickness of the emitter region 206 and the thickness of the HD collector region 707 may be substantially equal to the thickness of the base well in the vertical direction.
[0305] The drift region 205 of the L-BJT 1800 extends in a vertical direction from the top surface of the buried dielectric layer 909 to the bottom surface of the RESURF dielectric layer 734a. The drift region extension 304b of the L-BJT 1800 is bounded in a vertical direction by the capping dielectric layer 224 and the top surface of the buried dielectric layer 909. In some embodiments, the fully isolated L-BJT 1800 may not include a thick isolation dielectric layer above its drift region extension 304b. The field plate of the L-BJT 1800 includes a first field plate (RESURF layer) 732a, a second field plate 1812, and a second field plate 1818. In some cases, the second field plate 1812 may extend in a lateral direction from a first end substantially aligned with a middle region of the first field plate 732a to a second end above the drift region extension 304b. The third field plate 1818 may extend laterally away from the first field plate 1812 from a point near the second end of the first field plate 1812.
[0306] In some cases, the field plates 1812, 1818 of the L-BJT 1800 can keep the magnitude of the lateral electric field component in the drift region 304b constant or nearly constant from the emitter-base junction of the L-BJT 1800 to the HD collector region 707. Therefore, the V m Can be greater than 80 volts.
[0307] In some embodiments, the length of the extended drift region of the L-BJT 1800 can be further increased by increasing the lateral distance between the base well and the HD collector region 707 and adding a fourth field plate above the third field plate 1818 .
[0308] In some cases, the doping profile of L-BJT 1800 can be similar to the doping profile 1616 described with respect to L-BJT 350 .
[0309] In some embodiments, any of the L-BJTs described above (e.g., L-BJT 240, 730, 350, or 1600) may be isolated in a lateral direction using deep trenches filled with a dielectric material (e.g., oxide). These deep trenches may extend from the isolation dielectric layer 218 to a buried dielectric layer that vertically isolates (electrically isolates) an upper region of the substrate 290 where the L-BJT is formed below the doped well from a lower region of the substrate 290. Fig.19 An example HVL-BJT 1900 is schematically illustrated having an extended drift region that is fully isolated using a deep dielectric trench 1902 and a buried dielectric layer 1904. The fully isolated L-BJT 1900 includes one or more features described above with respect to the L-BJT 1600. The placement of the HD base region 806 of the L-BJT 1900 relative to the HD emitter region 206 is similar to the placement of the HD base region 806 of the L-BJT 804. In some examples, the height of the deep dielectric trench 1902 in the vertical direction (z-axis) can be 1 to 5 microns, 5 to 10 microns, 10 to 15 microns, 15 to 20 microns, or greater.
[0310] Fig. 20 The depletion region of L-BJT 2000 is schematically illustrated under three different magnitudes of bias or input voltages (e.g., voltages applied between HD collector region 707 and HD emitter region 206). The evolution of the depletion region of L-BJT 350, 1600, 1800, and similar L-BJT designs with extended drift regions may be at least qualitatively similar to the evolution of the depletion region of an L-BJT under the same bias conditions.
[0311] In some cases, the depletion region of L-BJT 2000 can be formed by two boundaries, including a first boundary near the emitter-base junction and the base well, and a second boundary surrounding the HD collector region 707 and extending to the collector region, sometimes extending to a portion of the substrate below the collector region.
[0312] As an example, at a relatively low bias voltage of about 5 volts, the first boundary 2102a may extend from the base region to the area below the base well, and the second boundary 2104a may extend along the entire drift region, including a portion of the substrate below the collector region.
[0313] As another example, at a moderate bias of about 80 volts, the first boundary 2102b may be confined in the well and surround the HD emitter region 206 and the HD base region 806, and the second boundary 2104b may shrink toward the HD collector region and may extend completely in the drift region.
[0314] As another example, at a relatively high bias voltage of about 250 volts, the first boundary 2102c may remain substantially unchanged compared to the first boundary 2102b at a bias voltage of 80 volts, while the second boundary 2104c may shrink further and be confined around the HD collector region.
[0315] While fully isolating the HV L-BJT using shallow or deep dielectric-filled trenches and buried dielectric layers based on the above architectures can be highly effective in preventing current and / or voltage coupling to low-voltage or medium-voltage devices fabricated on the same substrate, the implementation of these isolation structures can be costly and significantly increase the complexity and cycle time of the manufacturing process. Fabricating deep trenches and filling them with dielectric materials requires additional process steps that are longer and more complex than the fabrication of the local oxidation of silicon (LOCOS) and shallow trench isolation (STI) layers. The buried dielectric layer is typically embedded in the wafer or substrate on which the HV L-BJT is fabricated. An example of such a wafer is a silicon-on-insulator (SOI) wafer, which is much more expensive than a regular silicon wafer.
[0316] To reduce the cost, complexity, and cycle time of manufacturing an HV L-BJT that is properly isolated from low-voltage or medium-voltage devices fabricated on the same substrate, in some cases, the high-voltage region of the device can be isolated from other regions by a thick isolation dielectric layer on the surface of the device (e.g., by LOCOS fabrication). As described above, in an HV L-BJT, the high voltage supplied to the HD collector region is dropped along the extended drift region. Therefore, electrically isolating the HD collector region and the extended drift region from the base region and emitter region can decouple the high voltage from the substrate on which other electronic devices are fabricated.
[0317] Fig.21 Schematically shows a cross-sectional view of a HV L-BJT 2100 having an isolation drift region extension 2102 and four field plates 1510a, 150b, 1612, 1618. The architecture of the HV L-BJT 2100 is relatively similar to that of the HV L-BJT 1600 ( Fig. 16C ) is modified to allow the HD collector region 2101 and the drift region extension 2102 to be vertically isolated from the rest of the device using a second thick dielectric layer (e.g., a second field oxide layer) 2103, which operates at low or medium voltage levels. The HD collector region 2101 and the drift region extension 2102 have the same polarity, but the drift region extension 2102 can be a LD region with a smaller dopant concentration.
[0318] HV-LBJT 2100 may include emitter region 206, base well 208, spacer layer 216b, first field plate 1510a, second field plate 1510b, RESURF dielectric layer 734a, and thick isolation dielectric layer 1614. In some embodiments, the doping profiles, physical dimensions, configurations, and relative alignments of emitter region 206, base well 208, spacer layer 216b, first field plate 1510a, second field plate 1510b, RESURF dielectric layer 734a, and thick isolation dielectric layer 1614 of HV L-BJT 2100 may be the same as those of HV L-BJT 1600 ( Fig. 16C ) or 350( Fig.16A and 16B ) are the same or similar, and for the sake of brevity, their detailed descriptions are not repeated here.
[0319] The RESURF dielectric layer 734a, the thick isolation dielectric layer 1614, and the base well 208 can be formed in or disposed on a main well (e.g., a common well) 2114 having a polarity opposite to that of the drift region extension 2102. In some cases, the base well 208 can be partially formed in the main well 2114 and a second well 214 having a polarity opposite to that of the base well 208. In some cases, a vertical boundary between the second well 214 and the main well 2114 below the base well 208 can be aligned with a vertical edge between the emitter region and the base well below the spacer layer 216b.
[0320] In some cases, the second thick isolation dielectric layer 2103 is partially disposed or grown on the main well 2114 and the third well 2104 having the same polarity as the drift region extension 2102 (opposite polarity to the main well 2114).
[0321] The HV L-BJT 2100 also includes a third and fourth field plates 1612, 1618 extending laterally above a portion of the drift region extension 2102. In some cases, the third and fourth field plates 1612, 1618 are parallel to the top surface drift region extension 2102. The vertical distance h3 (e.g., along the z-axis) between the third field plate 1612 and the top surface of the drift region extension 2102 can be greater than the vertical distance h4 (e.g., along the z-axis) between the fourth field plate 1618 and the top surface of the drift region extension 2102. In various embodiments, the difference between h3 and h4 can be from 0.5 microns to 1.0 microns. In some examples, h1 and h2 can be 0.5 to 2.5 microns, or any value within a range defined by any of these values, or smaller or larger. In some cases, the third and fourth field plates 1612, 1618 can be manufactured in two different metallization layers (e.g., a metallization film of an IC) and can be electrically connected by vertical wires.
[0322] Similar to the HV L-BJT 1600, the third and fourth field plates 1612, 1618 can keep the lateral component of the E field in the drift region extension 2102 constant along the drift region extension 2102. The portion 2116 of the main well 2114 below the drift region extension 2102 can serve as an auxiliary field plate to further reduce the variation of the lateral component of the E field in the drift region extension 2102.
[0323] The third and fourth field plates 1612 , 1618 are electrically connected to the first and second field plates 1510 a , 1510 b , for example, via a conductive line 2112 having at least one lateral portion and one vertical portion.
[0324] In some cases, drift region extension 2102 of HV L-BJT 2100 may be electrically connected to main well 2114 via LD connection layer 2125 and HD connection region 2106 .
[0325] When the HV L-BJT 2100 is biased, for example, a potential difference is generated between the HD collector region 2101 and the emitter region 206, and the voltage variation of the device may have a stepwise behavior. Fig.21 The voltage curve shown in shows relatively constant voltage curve portions 2110a, 2110c and 2110e, and monotonically varying voltage curve portions 2110b, 2110d. As shown, the voltage remains relatively or nearly constant along the HD collector region 2101, as shown in voltage curve portion 2110e, varies (e.g., varies monotonically, e.g., varies substantially linearly) along the drift extension region 2102, as shown in voltage curve segment 2110d; the voltage remains relatively or nearly constant along the region extending from the boundary of the drift region extension 2102 to the edge of the second field plate 1510b, as shown in voltage curve portion 2110c, varies (e.g., varies monotonically, e.g., varies substantially linearly) along the drift region below the first and second field plates 1510a and 1510b, as shown in voltage curve segment 2110b; as shown in voltage curve portion 2110a, the voltage remains relatively or nearly constant in the base region. Therefore, the voltage along the drift region extends 2102 from the input voltage (V in ) level drops to the intermediate voltage (V Mid ), the drift region extends in a vertical direction and is electrically isolated from the substrate, and along the drift region, the voltage drops from the first voltage to the final voltage (V Fin In some cases, the first voltage may be equal to or less than 5 volts. In some cases, the ratio between the intermediate voltage and the input voltage (V Mid / V in ) can be less than 50%, less than 40%, less than 30%, less than 20%, or less than 10%. In some cases, by extending the drift region extension 2102, the voltage drop in the drift region extension (2102) can be increased, or VMid / V in To support a monotonic or substantially linear voltage drop along the drift region extension 2102, more field plates may be added above the drift region extension 2102. The vertical distance between each field plate and the top surface of the drift region extension 2102 may be smaller than the vertical distance between the subsequent field plate and the top surface of the drift region extension 2102.
[0326] In some cases, the emitter region 206, base well 208, spacer layer 216b, first field plate 1510a, second field plate 1510b, and thick isolation dielectric layer 1614 of the HV L-BJT 2100 and corresponding features of a MOS transistor (e.g., MOS transistor 280), an MV MOS or DMOS transistor (e.g., MOS transistor 740), or an HV MOS or DMMOS transistor (e.g., MOS transistor 300) can be manufactured together and have common physical dimensions.
[0327] In some embodiments, the length of the thick isolation dielectric layer 1614 in the lateral direction can be 0.5 to 5.0 microns, or any value within a range defined by any of these values or greater or less.
[0328] In some embodiments, the length of the first field plate 1510a (RESURF layer) can be from 0.18 microns to 1.0 microns, or any value within a range defined by any of these values, or larger or smaller.
[0329] In some embodiments, the length of the second field plate 1510b (RESURF layer) can be 0.5 to 5.0 microns, or any value within a range defined by any of these values or greater or less.
[0330] In some embodiments, the length of the third or fourth field plates 1612 , 1618 can be 1 to 15 microns, or any value within a range defined by any of these values, or larger or smaller.
[0331] In some examples, the lateral length of the third field plate 1612 may be 1-5 μm, and the lateral length of the fourth field plate may be 1-15 μm. Their vertical distances from the top surface of the drift region extension 2102 are 0.5 and 2.5 μm, respectively.
[0332] In some embodiments, the third and fourth field plates 1612, 1618 can cover 50% to 60%, 60% to 70%, 70% to 80%, 80% to 90% or 90% to 100% of the length of the thick isolation dielectric layer 1614, or any percentage within a range defined by any of these values or greater or less.
[0333] Although the region under the base well of the L-BJTs 350, 1600, and 2100 includes two regions with opposite polarities, in some cases, these regions under the base well may not include two regions with opposite polarities. As described above, dividing the region under the base well into two regions with opposite polarities and aligning the vertical interface between the two regions with the emitter-base junction can improve the performance of the L-BJT.
[0334] In various embodiments, an L-BJT (e.g., L-BJT 350 ( Fig.16A , 16B )、1600( Fig.18 )、1800( Fig.18 )) can be an NPN or PNP bipolar transistor. For example, an n-type emitter region, a p-type base well 208, an n-type extended drift region, and an n-type HD collector region can form an NPN bipolar transistor, and a p-type emitter region, an n-type base well 208, a p-type extended drift region, and a p-type HD collector region can form a PNP bipolar transistor. In some embodiments, an L-BJT with an extended drift region (e.g., L-BJT 350, 1600, 1800) can be at least partially co-fabricated with an HV MOS transistor (e.g., DMOS) or LV MOS transistor with an extended channel region. In some embodiments, one or more features of the HV L-BJT and the HV MOS transistor (or LV MOS transistor) can have common physical dimensions. For example, the bottom width of the spacer layer, the length of the RESURF layer and the RESURF dielectric layer, the length of the emitter region, the length of the base well, or the length of the HD collector region of the HV L-BJT may be substantially equal to the bottom width of the spacer layer, the gate layer and the gate dielectric layer, the HD source region, the LD source region, or the HD drain region of the HV MOS transistor.
[0335] In some embodiments, corresponding features of the HV MOS transistor and the HV L-BJT, such as those having common physical dimensions, can be cooperatively co-fabricated on a common substrate in the same fabrication steps. Unlike the HD and LD source regions of the HV MOS transistor, the emitter region and base well of the HV-LBJT are of different polarity, and in both cases the boundaries of the LD and HD regions are defined by the RESURF dielectric layer and the spacer layer.
[0336] As described above, the structure and configuration of the base well, base region, emitter region, RESURF layer, and RESURF stack in the HV L-BJT with extended drift region (e.g., HV L-BJT 350 or HV L-BJT 1600) can be the same as that of the LV L-BJT (e.g., L-BJT 240). Therefore, the above-mentioned manufacturing steps for co-manufacturing L-BJT 240 and MOST transistor 280 on a common substrate can be used to co-manufacture at least the base well, base region, emitter region, RESURF layer, and RESURF stack of HV L-BJT 350 (or L-BJT 1600) and the corresponding regions of MOS transistor 280.
[0337] Depletion Field Effect Divider
[0338] As mentioned above, about Figure 4A and 4B One aspect of the present invention relates to a structure for separating a low voltage (5V or less) region from a voltage drop region, such as in HV L-BJT 350 and HV MOS transistor 300. By physically separating the collector region along which the potential drops from a voltage greater than 5 volts to a current equal to or less than 5 volts, a wide range of voltages can be supported on a single chip, the chip including a first region on which LV MOS or LV BJT (e.g., L-BJT) devices are monolithically fabricated (e.g., co-fabricated or at least partially co-fabricated) and a second region in which a voltage divider (PD) is fabricated. In some cases, the first and second regions can be electrically isolated, and the devices in the first region can be electrically connected to the PD by wires, the wires including a vertical portion of the device connected to the PD and receiving a voltage from the PD, and a lateral portion connecting the vertical portions fabricated in a metallization layer above the first and second regions. The PD can be a silicon-based device and include features similar to the drift region extension of the HV L-BJT (e.g., HV L-BJT 350, 1600). A PD can be designed to monotonically or substantially linearly scale a high voltage (e.g., a voltage greater than 5 V, 10 V, 50 V, 80 V, 100 V, 200 V, or more) applied to an input port (or input node) to a low voltage (e.g., less than or equal to 5 V) at an output port (or output node).
[0339] In some cases, PDs can be manufactured using the same low-cost and fast cycle time bipolar CMOS (BiCMOS) core processes used to manufacture LV MOS and BJT transistors. This approach has the advantage of enabling higher voltage signals to scale monotonically or substantially linearly to the voltage nodes of any host process platform to enable a range of different design applications. Thus, advanced analog signal conversion between process platforms can be achieved.
[0340] In certain embodiments disclosed herein, a voltage divider (PD) may be a semiconductor device configured to operate based on the depletion field effect (DFE) principle. Therefore, the disclosed PD may be referred to as a depletion field effect PD or a DFE-PD. The DFE-PD monotonically or substantially linearly scales a higher input voltage at a first node (input node) to a lower voltage at a second node (output node). Therefore, the DFE-PD is advantageously suitable for analog signal conversion. In some embodiments, the DFE-PD is a semiconductor device comprising a plurality of individually doped silicon regions (e.g., four regions, or in some embodiments, five or more regions) that are arranged to form a first semiconductor junction at an input node and a second semiconductor junction at an output node. According to an embodiment, when the DFE-PD is biased (e.g., a voltage is applied to the input node), both the first and second junctions (e.g., PN junctions) may be reverse biased. When a voltage is applied to the input node, a portion of the charge induced in the input node may be coupled from the first junction to the second junction, generating an output potential in the output node, thereby acting as a voltage divider (PD). The ratio of voltage scaling may be defined by the ratio of the charge coupled from the first junction to the second junction. By electrically connecting the DFE-PD in series with an LV BJT, LV MOS transistor, or any suitable LV device, the voltage scaling operation can be electrically and physically decoupled from the low voltage BiCMOS platform.
[0341] In certain implementations, the full range of voltage scaling can range from a straightforward one-to-one voltage conversion to a situation where the output voltage is independent of the input voltage and fixed to be substantially constant. In some embodiments, the construction and operation of the depletion field device can exist between two extreme operating regions or device categories. At one extreme, a punch-through device is provided in which the charge applied to the first depletion region of the first junction is directly coupled to the second depletion region of the second junction so that the input voltage is substantially equal to the output voltage. At the other extreme, the charge provided to the input node is substantially completely coupled to one or more gate regions, which shield the output node from any charge coupling to the input node. This results in a device having a substantially fixed output voltage that is relatively independent of the input voltage.
[0342] It is desirable to create a DFE-PD device that operates between these two extremes so as to convert an input voltage signal monotonically or substantially linearly to a scaled-down output signal, at least within an input voltage range (e.g., a voltage range having a zero volt lower limit). To accomplish this, an architecture similar to a junction field effect transistor (JFET) or a modified JFET structure may be used in which the inter-gate region has a shortened length, photo-doping concentration, and the same polarity as the gate region (see the inter-gate region of a JFET, which has an opposite polarity to the gate region). In some cases, the inter-gate region on the DFE-PD is electrically connected to the gate region.
[0343] In some cases, shortening the length of the inter-gate region eliminates voltage saturation at the pinch-off output voltage, and the reverse polarity and low doping concentration of the channel region (compared to a JFET) produce a continuous voltage conversion curve with a linear or substantially linear region and shift the punch-through voltage of the device downward (e.g., to zero voltage). In some implementations, the length of the inter-gate region is balanced with the doping concentration so that the depletion region of the input junction (first junction) and the depletion region of the output junction (second junction) are coupled at zero input voltage. This results in a device that converts the input voltage monotonically or substantially linearly to a scaled-down output voltage, where the magnitude of the scaling factor is proportional to the ratio of charge coupling. Thus, the DFE-PD converts the input voltage (V in ) is scaled to a reduced output voltage (V out ), where the scaling ratio (V in -V out The slope of the curve can be adjusted to a suitable value by carefully adjusting the length of the gate inter-doping concentration.
[0344] Fig.22A A cross-sectional view of a semiconductor device 2200 for reducing voltage in the case of output saturation and a corresponding voltage transfer function in an open-circuit output configuration are schematically shown. The semiconductor device 2200 can monotonically or substantially linearly scale the input voltage until the output voltage reaches a saturation voltage (V p ). In some cases, the semiconductor device 2200 can be configured similar to a JFET. The semiconductor device 2200 includes an LD channel region 2205 extending from an input region 2202 to an output region 2204 along a first lateral or longitudinal direction (e.g., along an x-axis), and two highly doped (HD) gate regions 2206a, 2206b disposed on opposite sides of the LD channel region 2205 along a second lateral or longitudinal direction (e.g., a y-axis) perpendicular to the longitudinal direction. The input region 2202, the output region 2204, and the channel region 2205 have the same polarity opposite to the polarity of the gate regions 2206a, 2206b. For example, the gate regions 2206a, 2206b may include a p-type semiconductor, and the input region 2202, the output region 2204, and the channel region 2205 may include an n-type semiconductor. The first gate region 2206a and the second gate region 2206b may have the same nominal potential (e.g., may be electrically short-circuited). The first gate region 2206 a and the second gate region 2206 b form a PN junction 2208 with the channel region 2202 .
[0345] The carrier transport and E field distribution in the inter-gate region of the channel region 2205 are controlled by the electrical properties of the gate regions 2206a, 2206b and the potential applied to these gate regions. in -V outGraph 2209 shows the voltage difference between the output region 2202 and the HD gate regions 2206a, 2206b, and when the output region 2204 is open-circuited (e.g., R LOAD >1MΩ), the potential of the output region 2204 is a function of the potential of the input region 2202. As V in The increase of V out Monotonically or substantially linearly increases, and the gap between the depletion regions 2208 gradually becomes smaller. in =V c When the gap between the two depletion regions 2208 is closed, the electrical connection between the input region 2202 and the output region 2204 is pinched off by the depletion region, resulting in V out In V p (pinch-off voltage) saturates and pinches off.
[0346] Although the V in -V out behavior can be used to reduce and limit the voltage supplied to low-voltage devices, such as for protection from high-voltage damage, but for applications where high-voltage signals (such as time-varying signals) must be controlled and processed by circuits including LV devices, saturation of the output voltage may be caused by chopping voltages greater than V out Saturation voltage (V c ) to distort the signal. For such applications, the PD can be configured to monotonically or substantially linearly scale V by a scaling factor over a voltage range equal to or greater than the peak voltage of the high voltage signal that the LV device should handle. in Scale to V out .
[0347] By modifying the structure of semiconductor device 2200, monotonic or substantially linear scaling from Vin to Vout can be achieved. As a first step, the length of the inter-gate region of semiconductor device 2200 can be reduced to reduce V in -V out The saturation portion (V in >V c ) increases from zero to a positive value. Fig. 22B Schematically illustrates a cross-sectional view of another semiconductor device 2210 configured for voltage reduction without output saturation, and a corresponding voltage transfer function (V in -V out The structure of semiconductor device 2210 can be arranged to be the same as that of semiconductor device 2210, except that the length of the inter-gate region is shorter than that of device 2200 to allow some charge coupling from input region 2202 to output region 2204, even if V in Greater than V outIn some cases, the length of the inter-gate region of semiconductor device 2200 can be from 3.0 microns to 10 microns. As shown in curve 2219, when V out Less than V c When V out With the first slope V in increases monotonically or substantially linearly when V in Equal to or greater than V c When V out With a second slope smaller than the first slope, V in Therefore, since the inter-gate region is short even after pinch-off (V in >V c ), so the input to the inter-gate region (ΔQ in ) of the charge (ΔQ out ) is coupled to the output region 2204. Although V in =V c When , the gap between the two depletion regions 2218 is closed, but some electrical connection between the input region 2202 and the output region 2204 can still be supported by the charge tunneling pinch-off region. Given that the probability of charge tunneling is proportional to the length of the inter-gate region, the second slope can be adjusted by adjusting the length of the inter-gate region.
[0348] As V in -V out In a second step of substantially linearizing the relationship, the polarity of the inter-gate region of the semiconductor device 2210 can be configured to be the same as the polarity of the first and second gate regions 2216a, 2216b to electrically couple the first gate region 2216a to the second gate region 2216b, thereby in = 0 effectively results in a pinch-off between input region 2202 and output region 2204. This can reduce or eliminate V in -V out The first part of curve 2219 (V in <V c )( Fig. 22B ).
[0349] Fig. 22C A cross-sectional view of another semiconductor device 2220 for monotonic or substantially linear voltage reduction at a non-zero punch-through voltage (VPT≠0) and a corresponding voltage transfer function 2221 in an open-circuit output configuration are schematically shown. The inter-gate region 2213 of the semiconductor device 2220 has first and second gate regions 2216a, 2216b (for example, they can both be p-type regions). Therefore, the first gate region 2216a and the second gate region 2216b are electrically connected, and the charge coupling between the input region 2202 and the output region 2204 is V in= 0. However, since a new depletion region is formed between the gate inter-region 2213 and the input drift region 2212 and the output drift region 2214, the charge cannot be coupled from the input region 2202 to the output region 2204 until V in is large enough to close the gap between the two new depletion region boundaries 2228 (e.g., the minimum gap in the longitudinal direction) and allow charge to couple through the inter-gate region 2213 via a punch-through mechanism. in -V out As shown in Figure 2221, at the pinch-off voltage (V PT )2222, the charges are coupled through the inter-gate region 2213, V out With V out Linear or basically linear scaling, similar to V in -V out Curve 2219( Fig. 22B ) of the second region (V in >V c ), V in -V out The slope of the line is mainly determined by the length of the inter-gate region.
[0350] Make V in -V out The third step in linearizing the relationship is to reduce the dopant concentration of the inter-gate region of the semiconductor device 2220, thereby eliminating or otherwise significantly reducing the gap between the two depletion region boundaries 2228 to allow V in = 0 (or zero bias) through the inter-gate region charge coupling by the punch-through mechanism.
[0351] Fig.22D Schematically illustrates a cross-sectional view of an example DFE-PD device 2240 that can monotonically or substantially linearly convert V in Scale to V out , and the corresponding voltage transfer function 2241 in the open-circuit output configuration. The structure of DFE-PD 2240 is similar to that of semiconductor device 2220 ( Fig. 22C ) except that the majority carrier concentration in the inter-gate region 2215 is reduced (e.g., by reducing the doping concentration) compared to the device 2200, even if V in When V in -V out As shown in FIG. 2241 , the input voltage (V in ) can be scaled linearly or substantially linearly from V in =V out The output voltage (V out). and semiconductor device 2220 ( Fig. 22C ) is similar to V in -V out The slope is mainly determined by the length of the inter-gate region 2215. In some cases, the maximum voltage drop from the input region 2202 to the output region 2204 may be limited by the breakdown voltage of one of the junctions. In some cases, increasing the length (LD) of the input drift region 2212 may increase the DFE-PD according to its V in -V out The slope can be scaled down proportionally to the maximum input voltage. In some cases, increasing the length (L O ) can increase the maximum voltage that the DFE-PD can output.
[0352] Figures 23A-23B The evolution of the electric field and charge coupling through the inter-gate region 2215 of the example DFE-PD 2300 and the resulting voltage transformation are schematically and qualitatively shown. In the DFE-PD 2300, the input and output regions are eliminated, and the input region 2202 and the output region 2204 are directly in contact with the inter-gate region 2215. The majority carrier concentration in the inter-gate region 2215 and the length (L) of the inter-gate region 2215 are adjusted. p ), such that the depletion region boundaries 2228 touch or are separated by a small gap in the longitudinal direction (e.g., along the x-axis), wherein the gap can be less than 150 nanometers, 500 nanometers, 1000 nanometers, or a value between any of these values. When an input voltage (V in ), the charge (ΔQ in ) is coupled to the first and second gate regions 2216a, 2216b (at or substantially at the same potential), and ΔQ in The second portion (ΔQ2) of is coupled to the output region 2204, thereby generating an output voltage V between the first and second gate regions 2216a, 2216b at the output region 2204. out Therefore, with ΔQ in The first portion of the associated electric field is coupled to the first and second gate regions 2216a, 2216b (at or substantially at the same potential), and is related to ΔQ in The second portion of the associated electromagnetic field is coupled to the output region 2204, thereby generating an output voltage V between the first and second gate regions 2216a, 2216b at the output region 2204. out . With V in The increase of ΔQ1 and ΔQ2 and ΔQ in increases proportionally, making the ratio ΔQ2 / ΔQ inand the corresponding voltage ratio ΔV out / ΔV in The electric field is coupled to the first and second gate regions 2216a, 2216b, causing the output region 2204 to vary in the same manner as ΔQ1 and ΔQ2. in -V out As shown in Figure 2309, as V in Gradually increase from 0V, V out Increases with a slope less than 45 degrees (indicates a voltage decrease). V in -V out It shows that increasing V in Will not change V in -V out This indicates that the electric field coupling and charge coupling ratio are not affected by V in At low voltages, the point of contact between the depletion region boundaries 2228 (or the lateral location where the gap between them is reduced or minimized) moves toward the output region 2204, where the two boundaries eventually merge.
[0353] While the configuration of the DFE-PD 2300 simplifies the visualization of the electric field and charge coupling evolution, at least qualitatively, the DFE-PD 2240 ( Fig.22D )’s evolution of the electric field and charge coupler in the inter-gate region will be similar to that of the DFE-PD 2300.
[0354] Figures 24A-24B Two DFE PDs 2400, 2410 with different inter-gate regions are schematically shown, respectively. DFE-PD 2400 has a short inter-gate region 2415a, whose length (along the x-axis) is substantially less than its width (along the y-axis). In the example shown, the length of the inter-gate region 2415a is short enough that almost all of the E field lines and charges generated in the input region 2202 are coupled to the output region 2204. Therefore, V in -V out The slope of the line (in curve 2411) is close to 45 degrees, indicating that the voltage scaling ratio is close to 1 and the DFE-PD 2400 does not drop the voltage. In some examples, the length of the inter-gate region 2415a can be less than 5 microns, depending on the doping concentration of the input and output regions 2202, 2204 relative to the doping concentration of the inter-gate region 2415a.
[0355] The DFE-PD 2410 has a long inter-gate region 2415b, whose length (along the x-axis) is substantially greater than its width (along the y-axis). In the example shown, the length of the inter-gate region 2415a is long enough so that substantially all of the E field lines and charges generated in the input region 2202 are coupled to the first and second gate regions 2216a, 2216b. Thus, V in -V out The slope of the line (in curve 2412) is close to 0 degrees, indicating that DFE-PD 2410 generates a in Irrelevant constant (near zero) V out In some examples, the length of the inter-gate region 2415b is greater than 5.0 microns but less than 20.0 microns, depending on the doping concentration of the input and output regions 2202, 2204 relative to the inter-gate region 2415b.
[0356] Fig.25A Schematically shows a cross-sectional view of a DFE-PD 2500 with an extended drift region. Fig.22D ), the DFE-PD 2500 includes an input region 2202, an output region 2204, an inter-gate region 2215 extending in a longitudinal (e.g., x) direction from a first end to a second end, an input drift region 2212 extending from the input region 2202 to a first end of the inter-gate region 2215, and an output drift region 2214 extending from a second end of the inter-gate region 2215 to the output region 2204. The DFE-PD 2500 further includes a first gate region 2216a and a second gate region 2216b disposed above and below the inter-gate region 2215, respectively. The inter-gate region 2215 is laterally bounded by the first and second gate regions 2216a, 2216b extending from the first end to the second end of the inter-gate region 2215.
[0357] Fig.25B The values for eight different input voltages (V in ) values (including Vin = 40.5 volts (curve 2502), V in =35.5V (curve 2504), V in =30.5V (curve 2506), V in = 25.5 volts (curve 2508), V in = 20.5 volts (curve 2510), V in =15.5V (curve 2512), V in = 10.5V (curve 2514), V in=5.5 volts (curve 2516)), the calculated voltage change is plotted against the longitudinal distance from the output region 2204 of the DFE-PD 2500 (where 0 corresponds to the position output region 2204). The voltage at the x-axis value of -21 microns corresponds to V in (left vertical axis), the voltage at x-axis value 0 corresponds to V out (right vertical axis). As shown in graph 2502, the voltage is nearly constant along the input drift region 2212, decreases across the inter-gate region 2215, and is nearly constant along the output drift region 2214. Although the length (L p ) and the length of the output drift region 2214 (L o ) has no significant effect on the voltage drop across the DFE-PD 2500, but the maximum V that can be applied to the input region 2202 in and the maximum V that can be output by the DFE-PD 250 out , without damaging any junction or oxide layer, can be determined by L p and L o . Fig.26A shows that when the DFE-PD 2500 V in When changing from 0 to 40 volts, relative to V in Plotted calculation V out . V out -V in is not completely linear, but the average voltage scaling factor (V out / V in ) is about 5. Fig.26B shows the relative V in Plotted is the output current per unit length of the DFE-PD 2500. In some cases, V out -V in The linearity of the transfer curve is controlled by the geometry and doping profile of the inter-gate region 2215 relative to the doping profiles of the input drift region 2212 and the output drift region 2214. Optimizing these doping profiles and coupling geometries can improve linearity. Fig.26A and 26B The curves shown use a box-shaped doping profile to simplify calculations. Using graded doping profiles (along the x-axis, y-axis, or both) and carefully tailoring the doping profile can improve linear V out -V in curve (e.g., extending the linear region to a larger voltage range).
[0358] In some cases, increasing the length of the input drift region (L p ) and the distance between the output drift region (L o), and adding a RESURF plate (referred to as a “field plate” in this article) above it can increase the V that the DFE-PD can support. in and V out The field plates may function similarly to the field plates 1510a, 1510b, 1612, 1618 above the drift region extension 304b of the HV L-BJT 350, 1600 (FIG. 16). The field plates may maintain the lateral component of the E field in the input drift region 2212 and / or the output drift region 2214 substantially constant to generate a voltage drop along the input drift region 2212 and / or the output drift region 2214. Fig.27A A side cross-sectional view of a DFE-PD 2700 having gate regions 2716a, 2716b extending over the input drift region 2212 is schematically shown. The portion of the gate regions 2716a, 2716b extending over the input drift region 2212 can act as a field plate and produce a linear or substantially linear voltage drop 2702 along the input drift region 2212. In some cases, the length (L) of the portion of the gate regions 2716a, 2716b extending over the input drift region 2212 is 2.5 mm / s. resurf-1 ) can be smaller than the length (L D ). In various embodiments, a portion of the first gate region 2716a extending over the input drift region 2212 may include one or more electrically connected field plates (e.g., metal and / or polysilicon field plates) that overlap each other and are located above the input drift region 2212. These field plates may be connected to the portion of the gate region 2716a above the inter-gate region 2215. Similar to the field plates 1510a, 1510b, 1612, 1618 of the HV L-BJT 350, 1600, the vertical distance between these field plates and the top surface of the input drift region 2212 may increase from the interface between the integrating regions 2215 toward the input region 2202. In some examples, a portion of the well region or substrate in which the DFE-PD 2700 is formed may be electrically connected to the first gate region (e.g., through an ohmic contact) and serve as the second gate region 2716b and the one or more field plates.
[0359] In some cases, the maximum input voltage amplitude that the DFE-PD 2700 can scale down may be limited by the resulting reduced voltage, which may damage the junction and / or isolation dielectric layer (e.g., 5 volts) near the interface between the gate regions 2215 or within the output region 2204. For example, if the voltage scaling factor of the DFE-PD 2700 is 40, then V inThe upper limit of V may be 200 volts, since scaling any voltage above 200 volts by a factor of 40 may produce a voltage above 5 volts (e.g., 5 volts) near the interface between the inter-gate regions 2215 or within the output region 2204. In some embodiments, V may be increased by extending the output drift region 2214 and providing one or more mobile field plates above it. out upper limit (e.g., greater than 5 volts). Fig.27B A side cross-sectional view of a DFE-PD 2704 having gate regions 2717a, 2717b extending over the output drift region 2214 is schematically shown. The portion of the gate regions 2717a, 2717b extending over the output drift region 2214 can act as a field plate and produce a linear or nearly linear voltage drop along the output drift region 2214. In some cases, the length (L) of the portion of the gate regions 2717a, 2717b extending over the output drift region 2214 is 2.5 mm / s. resurf-2 ) can be smaller than the length (L O ). In various embodiments, a portion of the first gate region 2717a extending over the output drift region 2214 may include one or more electrically connected field plates (e.g., metal and / or polysilicon field plates) stacked on top of each other and located above the output drift region 2214. These field plates may be connected to the portion of the first gate region 2717a above the integrated region 2215. Similar to the field plates 1510a, 1510b, 1612, 1618 of the HV L-BJT 350, 1600 described above, the vertical distance between these field plates and the top surface of the output drift region 2214 may increase from the interface between the integrated regions 2215 toward the output region 2214. In some examples, a portion of the well region or substrate in which the DFE-PD 2704 is formed may be electrically connected to the first gate region 2717a (e.g., via an ohmic contact) and serve as the second gate region 2717b and the one or more field plates.
[0360] Fig.28A Schematically illustrated is a cross-sectional view of an actual implementation of a DFE-PD that can be fabricated with input / output regions, gate regions, and / or field plates on the same major surface of a substrate (e.g., the front side of a silicon substrate). The DFE-PD shown includes HD regions (2202, 2808, 2204, and 2812) and LD regions (2212, 2215, 2214) formed in a well 2828. The DFE-PD 2800 may also include an isolation dielectric layer (e.g., isolation dielectric layer 2818), and one or more field plates 2820 above the LD and HD regions. The process architecture of the DFE-PD 2800 is similar to that of the DFE-PD 2700 ( Fig.27A) process architecture, with similar electrical configurations for the LD and HD regions, but different physical geometries. The input drift region 2212 of the DFE-PD 2800 extends in a lateral direction (e.g., along the x-axis) from the input region 2022 to a first vertical interface with the inter-gate region 2215. The inter-gate region 2215 extends from the first vertical interface to a second interface with the output drift region 2214. The first vertical interface and the second vertical interface electrically correspond to the first and second junctions of the DFE-PD 2700 and the resulting depletion region 2228 ( Fig.27A ). DFE-PD2800 also includes an output region 2204, a first gate region 2808, and a gate contact region 2812 that provides electrical contact with a well (second gate region) 2828. The input drift region 2212, the inter-gate region 2215, and the region of the well 2828 below the output drift region 2214 serve as a similar circuit to that described above with respect to Fig.27B The second gate region 2717b of the DFE-PD 2704 is a second gate region.
[0361] In some cases, the input drift region 2212, the inter-gate region 2215, and the output drift region 2214 are doped regions formed in the well 2828 using an appropriate doping method (e.g., thermal diffusion or implantation) used in the MOS manufacturing process described herein. Similarly, the input region 2202, the output region 2204, the first gate region 2808, and the gate contact region 2812 are formed in the input drift region 2212, the inter-gate region 2215, and the output drift region 2214, respectively. In some cases, the input drift region 2212, the inter-gate region 2215, and the output drift region 2214 are LD regions with low doping concentration or low majority carrier concentration (e.g., lower than the HD region). In some cases, the input region 2202, the first gate region 2808, the output region 2204, and the gate contact region 2812 are HD regions with high doping concentration or high majority carrier concentration, and can be used as ohmic contacts.
[0362] The input region 2202, the input drift region 2212, the output drift region 2214, and the output region 2204 may have the same polarity. In some cases, the doping concentration or carrier concentration of the input region 2202 may be similar to or substantially the same as the doping concentration of the output region 2204. In some cases, the doping concentration or carrier concentration of the input drift region 2212 may be similar to or substantially the same as the doping concentration of the output drift region 2214.
[0363] First gate region 2808, inter-gate region 2215, gate contact 2812, and well 2828 may have the same dopant polarity type. In some cases, the doping concentration and / or majority carrier concentration of first gate region 2808 may be the same as the doping concentration or majority carrier density of gate contact 2812.
[0364] Still refer to Fig.28A , the DFE-PD 2800 also includes a thick isolation dielectric layer 2818 (e.g., an oxide layer such as a LOCOS or STI layer) located above the first drift region 2212, and three field plates 2820 disposed above the thick isolation dielectric layer 2818. The field plates 2820 may include a first field plate 2820a disposed on top of the thick isolation dielectric layer 2818, a second field plate 2820b placed above the first field plate 2810a and extending in a lateral direction toward the input region 2202, and a third field plate 2820c located above the second field plate 2820b and extending in a longitudinal direction toward the input region 2202. In some cases, the first, second, and third field plates 2820a, 2820b, 2820c may cover a length (L) of the input drift region 2212. D ) part (L resurf-1 ). In some examples, L D Can be compared to L resurf-1 1.2, 1.4, 1.6, 1.7, 2 times or more or less.
[0365] In some cases, the arrangement of the first, second, and third field plates 2820a, 2820b, 2820c can be similar to the arrangement of the second, third, and fourth field plates 1510a, 1612, 1618 described above with respect to the HV L-BJT 1600. In some cases, the first, second, and third field plates 2820a, 2820b, 2820c can include features (e.g., geometry and material properties) similar to the second, third, and fourth field plates 1510a, 1612, 1618, respectively. For example, the first field plate 2820a can include polysilicon, and the second field plate 2820b and the third field plate 2820c can include a metal or a metal alloy. Similar to the second, third, and fourth field plates 1510a, 1612, 1618 of the HV L-BJT 1600, the first, second, and third field plates 2820a, 2820b, 2820c can keep at least the lateral component of the E field in the input drift region 2212 constant or nearly constant, for example by introducing a vertical E field component (e.g., along the y-axis) along the input drift region 2212. As a result, in addition to the voltage drop (ΔV dep ), the presence of the field plate may also cause a voltage drop (ΔV drift The final voltage scaling factor of DFE-PD 2800 can be greater than the voltage scaling factor of DFE-PD 2500 ( Fig.25A ).
[0366] In various implementations, the input drift region 2212 may be varied by changing its length (L D ) to adjust ΔVdrift , by changing the length of the gate region 2808 (L p ) and the majority carrier concentration (or doping concentration) in the gate inter-region 2808 to adjust ΔV dep In various embodiments, ΔV dep The voltage range is 0 to 5 volts, ΔV drift The voltage range is also 0 to 300 volts.
[0367] In various embodiments, L p It can be 0.1 to 5.0 microns, L D The gate area 2808 may be 0.5 to 25 microns, and the majority carrier concentration (or doping concentration) in the gate area 2808 may be 10 15 Up to 10 17 cm -3 In some examples, the majority carrier concentration (or doping concentration) in the inter-gate region 2808 is lower than the majority carrier concentration in the first gate region 2808 and the well (second gate region) 2828. In some embodiments, the difference between the majority carrier concentration (or doping concentration) in the inter-gate region 2808 and the first gate region 2808 can be from 10 16 to 10 17 cm -3 In some examples, the difference between the majority carrier concentration (or doping concentration) in the inter-gate region 2808 and the well (second gate region) 2828 may be from 10 16 to 10 17 cm -3 .
[0368] In some cases, field plate 2820 and gate contact 2812 can be at substantially the same potential. For example, during operation of DFE-PD 2800, they can be connected to an ohmic region or electrode (eg, a ground plane) that has a stable and constant potential.
[0369] Curve 2802 shows the input voltage (V in ) for four different values, including V in =200V (curve 2830), V in =144V (curve 2832), V in = 100 volts (curve 2834) and V in = 2836 volts (curve 2836), the calculated change in voltage along the lateral direction (e.g., along the x-axis) of DFE-PD 2800 plotted against the distance from input region 2202. Graph 2803 is a close-up view of a portion of graph 2802 near the interface between inter-gate region 2215 and within output region 2204.
[0370] In some embodiments, fabrication of the DFE-PD 2800 may include forming a doped substrate region 2828 by implanting the top layer of a silicon-on-insulator (SOI) wafer and thermally driving dopants downward toward the buried oxide layer 1904. Next, an epitaxial layer is grown on top of the silicon layer and a single well is implanted the entire length from region 2214 to 2212. Next, a second well of opposite dopants is implanted to form the inter-gate region 2215. In some cases, the dopant concentration of this well will exceed the dopants of the well regions 2214 and 2212. These wells will be thermally driven until they merge with the buried dopant layer 2828. The excess dopant of 2215 will counter-dope 2214 and 2212 to form a lighter net doped region 2215. Next, a dielectric layer is formed and heavily doped regions (ohmic contacts) 2202, 2208, 2204, and 2812 are implanted. Finally, metal interconnects are formed.
[0371] Fig.28B shows that when the V in When V changes from 0 to 200 volts, out Relative to V in Example calculations plotted. V out -V in is not completely linear, but the average voltage scaling factor (V out / V in ) is about 40. Fig.28C shows the relative V in An example calculation of the output current per unit length of the DFE-PD 2800 is plotted.
[0372] Fig.28D Snapshots of the depletion region of the DFE-PD 2800 are schematically shown at five different magnitudes of bias or input voltage (e.g., voltage applied between the input region 2202 and the first gate region 2808). Associated with each junction are two depletion boundaries that push in on either side of the opposite junction dopant. Fig.28D In FIG. 1 , depletion boundaries 2820a-2820e and 2824a-2824e represent one side of the eth depletion regions of the two junctions. The depletion regions on the other side of the junctions in the inter-gate region are merged. In some cases, the depletion boundaries represented by 2822a-2822e show how the merged depletion between gates cannot merge around the higher doped regions.
[0373] As described above, in some applications, the high input voltage can be scaled down to a lower voltage, but it may still be too high (e.g., 5 volts) for the junctions and dielectric layers near the interface between the gate regions 2215 or within the output region 2204. For example, a region or device in an IC may be designed to operate at 50 volts, while the available power supply can produce an output voltage of 200 volts (voltage scaling factor of 4). Although by reducing L p and / or L D , the voltage scaling factor of the DFE-PD 2800 can be reduced to 4, but the output region 2204 may be damaged by the resulting scaled voltage (50 volts). This limitation can be eliminated by adopting a longer output drift region 2204 and adding one or more field plates (RESURF plates) above the output drift region 2204 (the same design as used for the input drift region 2212).
[0374] Fig.29 Schematically illustrated is a cross-sectional view of a DFE-PD 2900 having an extended output region 2914 and two field plates 2944, 2942 placed above the extended output region 2914. In some examples, the input region, output drift region, dielectric layer on the input region, gate interlayer, gate layer, and well forming the DFE-PD 2900 are similar to corresponding features of the DFE-PD 2800.
[0375] Unlike DFE-PD 2800, DFE-PD 2900 includes an extended output region 2914 and a thick dielectric layer 2940 disposed thereon. In some cases, field plates 2944, 2942 may cover the length (L) of the extended output drift region 2914. O ) part (L resurf-2 ). In some examples, L O Can be compared to L resurf-2 1.2, 1.4, 1.6, 1.7, 2 times greater, or having a value within a range defined by any of these values or a greater or lesser value. Extending output region 2914 and field plates 2944, 2942 can allow DFE-PD 2900 to support output voltages higher than 5 volts. The upper limit of the output voltage amplitude of DFE-PD 2900 can be increased by increasing L O and L resurf-2 To increase.
[0376] In some cases, field plates 2944, 2942 and gate contact 2812 can be substantially at the same potential. For example, during operation of DFE-PD 2900, they can be connected to an ohmic region or an electrode with a stable and constant potential (e.g., a ground plane).
[0377] Thick dielectric layers 2818 and 2940 may include one or more features described above with respect to thick dielectric film 1614 of HV L-BJT 350 ( Fig.16A , 16B ).
[0378] To support a voltage drop (e.g., a substantially linear voltage drop) along the input drift region 2112 and / or the output drift region 2214, more field plates may be added above each drift region. The vertical distance of each field plate from the top surface of the corresponding drift region may be less than the vertical distance of the subsequent field plate. The length of each field plate and its relative vertical distance relative to the adjacent field plate may be determined based on the input voltage or voltage range of the PD and the desired output voltage or voltage recovery provided by the PD.
[0379] In some embodiments, DFE-PD 2800 or DFE-PD 2900 may be isolated in a lateral direction using deep trenches 1902 filled with a dielectric material (e.g., oxide). These deep trenches may extend from the isolation dielectric layer 218 to a buried dielectric layer 1904 that vertically isolates (electrically isolates) an upper region of the substrate where the well 2828 is formed (e.g., below the well 2828) from a lower region of the substrate.
[0380] MOS device coupled with high voltage divider
[0381] As described above, another second inventive aspect allows a low voltage BiCMOS platform to be boosted to a higher voltage node using a high voltage voltage divider connected in series with a low voltage device to create a high voltage device.
[0382] In some embodiments, the above-described DFE PD can be electrically connected (e.g., in series) to a low voltage or medium voltage bipolar junction device (e.g., L-BJT) or a low voltage MOS device (e.g., MOSFET). In some cases, the DFE PD, bipolar device, and MOS device can be manufactured on a common substrate using MOS manufacturing techniques and libraries similar to those described above. In some such cases, at least a portion of the DFE-PD can be co-fabricated with the bipolar and / or MOS devices. The DFE PD can be laterally separated or isolated from the bipolar device and the MOS device, and / or vertically separated or isolated from the common substrate.
[0383] In some cases, the DFE PD can be fabricated on an integrated circuit (IC) including bipolar junction devices and MOS devices. In some cases, the DFE PD can be fabricated on a high voltage region of the IC that is electrically isolated from a low voltage region of the IC (e.g., by a deep dielectric filled trench). In some examples, the DFE PD can be connected to devices in the low voltage region through vertical wires (vias) that are connected to common lateral wires fabricated, for example, in a metallization layer of the IC above a substrate. In some embodiments, the DFE-PD can be electrically connected to multiple low voltage devices that receive a reduced voltage from the DFE-PD. In various embodiments, the DFE-PD can have a voltage scaling factor of 2 to 5, 5 to 10, 10 to 20, 20 to 30, 30 to 40, 40 to 50, or a value within a range defined by any one of these values or greater. The DFE-PD can be configured to reduce a high voltage level to a voltage level that is safe for a low voltage device (e.g., equal to or less than 5 volts) or a medium voltage device (e.g., equal to or less than 80 volts).
[0384] In some embodiments, the DFE-PD and the low voltage device (eg, L-BJT or MOS transistor) may have at least one layer having a common material or physical dimensions, or at least one doping region having a common doping profile.
[0385] The input region, output region, and drift region of the DFE-PD may have the same polarity as the drain region of the MOS transistor or the collector region of the L-BJT to which they are connected.
[0386] In a similar manner to the co-processing advantages of HV L-BJTs described above, various co-processing of DFE PD with L-BJTs and MOS transistors can be performed in parallel, including defining one or more thick dielectric layers, field plates, extended drift regions, and HD and LD doped regions, respectively. The details of the co-processing have been described above and will not be repeated here for the sake of brevity.
[0387] In some cases, the DFE-PD and the low voltage (LV) MOS transistor that receives the reduced voltage from the DFE-PD can be electrically similar to a single HV MOS transistor, such as an LDMOS transistor. The channel region of such an LDMOS transistor can be laterally inserted between the HD input region 2202 of the DFE-PD serving as the eth LDMOS drain and the HD source region of the LVMOS having the same polarity as the HD input region 2202. The HD output region 2204 of the DFE-PD is physically separated from the HD drain region of the LV MOS by an inter-gate structure while being electrically connected to the HD drain region of the LV MOS. The LDMOS transistor may also include a back gate contact electrically connected to a field plate 2820 formed on an extended input drift region serving as a reduced surface field (RESURF) plate. The LDMOS transistor may include a back gate contact electrically connected to the gate region 2802 of the DFE-PD.
[0388] Fig. 30A Schematically illustrates a cross-sectional view of a DFE-PD 2800 that is similar to, for example, the L-BJT 240 ( Figure 2 ) of the lateral BJTs. The output region 2204 of the DFE-PD 2800 is connected to the HD collector region 707 of the L-BJT 240 via the first conductive line 3000. The conductive line 3002 may be a conductive line patterned in a metallization layer above the substrate on which the DFE-PD 2800 and the L-BJT 240 are fabricated, and the output region 2204 and the HD collector region 707 may be electrically connected to the first conductive line 3000 through two separate vertical conductive lines (vias).
[0389] The first gate region 2808, field plate 2820, and gate contact 2812 of DFE-PD 2800 are connected to RESURF contact 712b and RESURF layer 714b of L-BJT 240 via second conductive line 3002. Second conductive line 3002 may be a conductive line in a metallization layer above a substrate on which DFE-PD 2800 and L-BJT 240 are fabricated, and first gate region 2808, field plate 2820, gate contact 2812, and RESURF contact 712b may be electrically connected to second conductive line 3002 through two separate vertical conductive lines (e.g., vias). In various embodiments, first and second conductive lines 3000, 3002 may be fabricated on the same or different metallization layers.
[0390] Fig. 30B Schematically illustrates a side cross-sectional view of a DFE-PD 2800 that is similar to, for example, the MOS transistor 280 described above ( Figure 2) are laterally isolated and / or separated and electrically connected to the low voltage MOS transistors of the DFE-PD 2800. The output region 2204 of the DFE-PD 2800 is connected to the HD drain region 721 of the MOS transistor 280 via a third conductive line 3004. The third conductive line 3004 may be a conductive line patterned in a metallization layer above a substrate on which the DFE-PD 2800 and the MOS transistor 280 are fabricated, and the output region 2204 and the HD drain region 721 may be electrically connected to the third conductive line 3004 via two separate vertical conductive lines (e.g., vias).
[0391] The first gate region 2808, the field plate 2820, and the gate contact 2812 of the DFE-PD 2800 are connected to the gate contact 712a and the gate layer 714b of the MOS transistor 280 via the fourth conductive line 3006. The fourth conductive line 3006 may be a conductive line patterned in a metallization layer above the substrate on which the DFE-PD 2800 and the MOS transistor 280 are fabricated, and the first gate region 2808, the field plate 2820, the gate contact 2812, and the gate contact 712a may be electrically connected to the fourth conductive line 3006 by two separate vertical conductive lines (e.g., vias). In various embodiments, the first, second, third, and fourth conductive lines 3000, 3002, 3004, 3006 may be fabricated on the same or different metallization layers.
[0392] In some cases, one or both of MOS transistor 280 and L-BJT 240 may be fabricated on the same substrate as DFE-PD 2800. For example, MOS transistor 280 and L-BJT 240 may be fabricated on a low voltage region of an IC that is electrically isolated from a high voltage region of the IC on which DFE-PD 2800 is fabricated. Figures 30A-30B As shown, in some cases, electrical isolation may be provided by deep dielectric-filled trenches disposed between low voltage devices (eg, MOS transistor 280, L-BJT 240) and DFE-PD 2800 and / or between low voltage and high voltage regions of an IC.
[0393] In some cases, the HD gate region 721, the HD source region of the MOS transistor 280, and the collector region 707 of the L-BJT 240 can be fabricated together with the output regions 2202, 2204 of the DFE-PD 2800. High voltage devices, including voltage dividers physically separated from low voltage devices
[0394] In some embodiments, the above-mentioned DFE-PD (e.g., DFE-PD 2900 ( Fig.29 ) or 2800( Fig.28A)) can be physically separated and / or electrically isolated from one or more low voltage or medium voltage devices that receive scaled-down voltages from the DFE-PD to avoid any damage to the low voltage or medium voltage devices by the high voltage signals. In various embodiments, the physical and / or electrical isolation can include lateral and / or vertical separation from the low voltage or medium voltage devices via one or more dielectric layers. In some cases, lateral isolation may include isolation through shallow or deep dielectric-filled trenches formed in the substrate or through an oxide layer (e.g., LOCOS) thermally grown on the top (main) surface of the substrate. In some cases, vertical isolation can be provided by a buried oxide layer, such as silicon on an insulator substrate. In some other cases, vertical isolation can be provided by a thick dielectric layer (e.g., a thermally grown or deposited oxide layer) formed on or above the substrate and / or by a portion of one or more interlayer dielectric layers (ILD) or intermetallic dielectric layers (IMD) deposited above the substrate. In some cases, the cladding layer 224 may include one or more ILD or IMD layers.
[0395] Fig.31 Schematically illustrated is a cross-sectional view of an example high voltage device including electrical connections between a DFE-PD 2800 and a low voltage device such as a LV L-BJT 804 fabricated on a common substrate for controlling (e.g., switching, modulating, etc.) a high voltage signal provided to the DFE-PD 2800 using the low voltage device (e.g., LV L-BJT 804). The DFE-PD 2800 and the L-BJT 804e are laterally separated or isolated from each other using deep dielectric-filled trenches 1902 and buried dielectric layers 1904. In some cases, the substrate may include a silicon-on-insulator (SOI) wafer, where the DFE-PD 2800 and the LV L-BJT 804 are fabricated on a top silicon layer above an insulating layer (e.g., a buried oxide layer of the silicon-on-insulator). The buried dielectric layer 1904 can provide advantages, including, for example, reducing the depth of the trench 1904 by blocking carriers from flowing between the two devices through unblocked areas of the substrate below the trench 1902. Fig.31 As shown, similar dielectric trenches can be used to electrically isolate DFE-PD 2800 and LV L-BJT 804 from other devices fabricated on a common substrate. Deep trench 1902 can extend from isolation dielectric layer 218 to buried dielectric layer 1904. In some cases, deep trench 1902 can contact buried dielectric layer 1904. In some cases, the vertical gap between deep trench 1902 and buried dielectric layer 1904 can be less than 50 nm, or less than 10 nm.
[0396] The output region 2204 and the HD collector region 707 of the LV BJT 804 are electrically connected via a first conductive line 3000, which is fabricated in a metallization layer connected above the top surface of the common substrate. The field plate 2820, the first gate region 2808, the gate contact region 2812, and the RESURF layer 714b are electrically connected by a second conductive line 3002 fabricated in a metallization layer above the top surface of the common substrate. The first conductive line 3000 and the second conductive line 3002 may be fabricated on the same or different metallization layers.
[0397] In various implementations, the inter-gate region of the DFE-PD can control the coupling ratio between the input voltage and the corresponding output voltage in two configurations; the first configuration described above includes one or more interfaces or junctions with HD gate regions on the top or bottom or both sides of the inter-gate region, and the electric field coupling with these gate regions controls the voltage drop across the inter-gate region. In the second configuration, the DFE-PD can control the voltage drop across the inter-gate region based on electric field coupling with one or more adjacent HD regions of electrodes that are not in contact with the inter-gate region. The electric field coupled to such electrically isolated HD regions or electrodes can have the same effect on the voltage drop across the inter-gate region and control the ratio between the input and output voltages as the HD gate regions 2216a and 2216b.
[0398] In some embodiments, the architecture of DFE-PD 2800 can be modified by removing one or both of gate regions 2216a, 2216b and relying on the dynamics of depletion region boundary 2228 for voltage scaling. In some cases, second gate region 2216b can be eliminated, and first gate region 2216a can be separated from inter-gate region 2215. In these examples, the region where the dielectric region is located, also referred to herein as inter-gate region 2215, can be referred to as a depletion control region.
[0399] Fig.32A cross-sectional view of an example high voltage device is schematically shown, the example high voltage device including an electrical connection between a DFE-PD 3206 and an L-BJT 3208 fabricated on a common substrate, wherein the DFE-PD 32006 and the L-BJT 3208 (similar to the L-BJT 800) are laterally separated or isolated from each other by a dielectric layer 218, and vertically separated or isolated from a lower region of the common substrate by a buried dielectric layer 909. The dielectric layer 218 can be a shallow dielectric field trench (e.g., a shallow trench oxide) or a thermally grown oxide layer. In some examples, the buried dielectric layer 909 can be in contact with or merged with the dielectric layer 218. In some cases, the thickness of the dielectric layer 218 in the vertical direction can be 0.3 to 0.5 microns. In some examples, the vertical gap or distance between the bottom surface of the dielectric layer 218 and the top surface of the buried dielectric layer 909 (e.g., along the z-axis) can be less than 5nm, or less than 10nm. DFE-PD 3206 may include input region 2202, input drift region 3212, depletion control region 3220, and output region 2204. The polarity and relative majority carrier concentration (or doping concentration) in these regions may be similar to DFE-PD 2800 ( Fig.28A) of the input region 2202, the input drift region 2212, the inter-gate region 2215, and the output region 2204. The DFE-PD 3206 may also include a thin dielectric layer 3203 located above the depletion control region 3220, a first field plate layer 2820a, a spacer layer 3204 formed on the sidewalls of the first field plate 2820a, a second field plate 2820b located above the first field plate 2820a, and a third field plate 2820c located above the second field plate 2820b. The input drift region 3212 is vertically bounded by one or more inter-metal dielectric layers 224 and the top surface of the buried dielectric layer 909. In some embodiments, the DFE-PD 3206 may not include a thick isolation dielectric layer above its input drift region 3212. In some cases, the second field plate 2820b can be connected to the first field plate 2820a by a vertical connector (e.g., a through hole) and extends in a lateral direction from a first end substantially aligned with the midpoint of the first field plate 2820a to a second end above the input drift region 3212. The third field plate 2820c can be connected to the second field plate by a vertical connector (e.g., via) and extends in a lateral direction from an area near the second end of the second field plate 2820b. The field plate 2820 of the DFE-PD 3206 is configured to keep the amplitude of the lateral electric field component in the input drift region 3212 relatively or nearly constant from the input region 2202 to the output region 2204. The output region 2204 and the HD collector region 207 of the LV BJT 3208 are electrically connected by a first conductive line 3000 fabricated in a metallization layer above the top surface of the common substrate. The field plate 2820 is electrically connected to the RESURF layer 714b via a second conductive line 3002 fabricated in a metallization layer above the top surface of the common substrate. The first and second conductive lines 3000, 3002 may be fabricated as part of the same or different metallization layers.
[0400] Fig.33A cross-sectional view of an example high voltage device including a DFE-PD that is vertically separated from a substrate and directly connected to an L-BJT is schematically shown. Device 3300 includes a DFE-PD 3320 that is directly connected to an L-BJT 730 while being vertically elevated or separated from the substrate on which the L-BJT 730 is fabricated. The DFE-PD 3320 and the L-BJT 730 are fabricated on a common substrate, thereby forming an integrated device that effectively functions as a HV L-BJT. The DFE-PD 3320 includes an input region 2202, an input drift region 3312, and a depletion control region 3220 on a top surface of a thick dielectric layer 3304 disposed on a common substrate. In some cases, advantageously, the input region 2202, the input drift region 3312, and the depletion control region 3220 can be formed from the same layer, such as a polysilicon layer. In these embodiments, the thickness of these layers in the vertical direction (along the y-axis) can be substantially equal. In some other cases, a base layer, such as a polysilicon layer, which may form the input region 2202, the input drift region 3312, and the depletion control region 3220, may be co-deposited and co-patterned to have the same thickness as the RESURF layer 1510, so that the thickness of all these regions along the vertical direction is substantially equal. In addition, in some cases, one or more of the input region 2202, the input drift region 3312, and the depletion control region 3220 may be co-doped to have the same dopant type and / or concentration as the RESURF layer 1510. In various embodiments, the thickness of the input region 2202, the input drift region 3312, and the depletion control region 3220 may be 0.15 to 0.5 μm.
[0401] The DFE-PD 3320 includes an output drift region 2214 and an output region 2204. The input drift region 3312 extends from the input region 2202 to the depletion control region 3220 in a lateral direction. The output drift region extends from the depletion control region 3220 to the output region 2204. The output region 2204 is formed on or partially in the first substrate region 3302. The first region of the substrate 3302, the output region 2204, the input drift region 3312 and the input region 2202 of the DFE-PD 3320, and the emitter region 206 of the L-BJT 730 may have the same polarity opposite to the polarity of the depletion control region 3220 of the DFE-PD 3320 and the base region 208 of the L-BJT 730. The output region 2204 of the DFE-PD 3320 may extend laterally from the thick dielectric layer 3304 and the thick dielectric film 734b disposed on the drift region of the L-BJT 730. In some examples, the output region 2204 of the DFE-PD 3320 may serve as a HD collector region of the L-BJT 730. Thus, the output region 2204 serves as a common electrical contact between the DFE-PD 3320 and the L-BJT 730 that are electrically arranged in series. In some cases, the combination of the DFE-PD 3320 and the L-BJT 730 (the device within the dashed box 3340) may electrically behave as a HV L-BJT, with the input region 2202 serving as an effective collector region of the HV L-BJT.
[0402] The DFE-PD 3320 may also include two electrically connected field plates 2820b, 2820c located above the input drift region 3312, extending from the interface between the input drift region 3312 and the depletion control region 3220 toward the input region 2202. Each field of the plates 2820b, 2820c may be fabricated as part of a different metallization layer, wherein the field plate 2820 closer to the depletion control region 3220 has a smaller vertical distance from the top surface of the input drift region 3312 than the other field plate closer to the input region 2202.
[0403] The common substrate may include a second substrate region 214 having a polarity opposite to that of the first substrate region 3302. The base well 208 of the L-BJT 730 may have a portion formed in each of the first substrate region 3302 and the second substrate region 214 and extend across a boundary therebetween.
[0404] In some cases, the first substrate region 3302 below the thick dielectric layer 3304 serves as a bottom field plate for the DFE-PD 3320 and helps keep the lateral component of the E field within the input drift region 3312 relatively constant or nearly constant (this situation results in a monotonic or substantially linear decrease in voltage along the input drift region 3312).
[0405] Field plates 2820b, 2820c and RESURF layer 1510b can be electrically connected by wires 3002 fabricated in a metallization layer above the common substrate top surface. Wires 3002 can be fabricated on the same metallization layer on which field plates 2820b or 2820c are fabricated, or on different metallization layers.
[0406] refer to Fig.33 The voltage and distance curve shown in the lower part shows that when the input voltage (V in ), the voltage variation along the device may have a segmented behavior, and the voltage-to-distance curve may include substantially constant segments 3310a and 3310d, and decreasing segments 3310b, 3310c, and 3310e. For example, the voltage may remain constant or nearly constant in segment 3310a along the input region 2202, decrease (e.g., monotonically or substantially linearly) in segment 3310b along the input drift region 3312, decrease (e.g., monotonically or substantially linearly) in segment 3310c along the depletion control region 3220 at the same or different slope as segment 3310b, the voltage may remain constant or nearly constant in segment 3310d along the output region 2204 and the first portion of the drift region of the L-BJT 730, decrease (e.g., monotonically or substantially linearly) in segment 3310e along the second portion of the drift region below the RESURF layer 1510, and the voltage may remain constant 3310e along the base region of the L-BJT 730. Thus, the voltage decreases from the input voltage (V ) applied to the input region 2202 along the first portion of the drift region. in ) level drops to the intermediate voltage (V Mid ), along the second part of the drift region from the intermediate voltage to the final voltage (V Fin ). In some cases, V Fin can be equal to or less than 5 V. In some cases, the ratio between the intermediate voltage and the input voltage (V Mid / V in ) may be less than 50%, less than 40%, less than 30%, less than 20%, or less than 10%. In some cases, the slopes of the voltage drop segments 3310b, 3310c, 3310e may be different and may be individually customized by adjusting the length of the input drift region 3312, the length and / or doping of the depletion control region 3220, and the length of the drift region of the L-BJT 730, respectively.
[0407] Fig.34A cross-sectional view of an example high voltage device including a DFE-PD vertically separated from a substrate and electrically connected to an L-BJT via a wire is schematically shown. The device 3400 includes a DFE-PD 3420 electrically connected to a LV L-BJT 3410 while being vertically isolated and / or separated from the substrate on which the LV L-3410 is fabricated. The DFE-PD 3420 and the LV L-BJT 3410 are fabricated on a common substrate 290, thereby forming an integrated device 3400 that effectively functions as a HV L-BJT. The DFE-PD 3420 includes an input region 2202, an input drift region 3312, a depletion control region 3220, an output drift region 2204, and an output region 2204 disposed on a top surface of a thick dielectric layer 3304 on the common substrate 290. The input drift region 3312 extends in a lateral direction from the input region 2202 to the depletion control region 3220 , and the output drift region 2214 extends laterally from the depletion control region 3220 to the output region 2204 .
[0408] The DFE-PD 3420 may also include two electrically connected field plates 2820b, 2820c located above the input drift region 3312, extending from the interface between the input drift region 3312 and the depletion control region 3220 toward the input region 2202. Each field of the plates 2820b, 2820c may be patterned on a different metallization layer, wherein the field plate 2820 closer to the depletion control region 3220 has a smaller vertical distance from the top surface of the input drift region 3312 than the other field plate closer to the input region 2202.
[0409] The output region 2204 of the DFE-PD 3420 and the HD collector region of the LV L-BJT 3410 may be electrically connected via a first conductive line 3000 fabricated in a metallization layer above the common substrate top surface. The RESURF layer 714b of the LV L-BJT 3410 may be electrically connected to the field plate 2820b via a second conductive line 3002 fabricated in a metallization layer above the common substrate top surface. The first conductive line 3000 and the second conductive line 3002 may be fabricated on the same or different metallization layers.
[0410] In some cases, the DFE-PD 3320 and the L-BJT 730 can be fabricated in a well 3402 formed in a common substrate 290. The well 3402, the input region 2202, the output region 2204, the input drift region 3312, the HD collector region 207, and the emitter region 206 can have the same polarity opposite to that of the base well 208 and the depletion control region 3220. In some cases, the majority carrier concentration in the depletion control region 3220 can be less than the majority carrier density of the input and output drift regions 3312, 2214.
[0411] In some cases, the combination of DFE-PD 3320 and L-BJT 730 (device within dashed box 3440 ) may electrically behave as a HV L-BJT, where input region 2202 serves as an effective collector region of the HV L-BJT.
[0412] Fig.35 A cross-sectional view of an example high voltage device including a DFE-PD that is laterally and vertically separated and / or isolated from a substrate and electrically connected to an L-BJT by a conductor is schematically shown. The device includes a DFE-PD 3520 that is laterally and vertically separated and / or isolated from a substrate and electrically connected to an LV L-BJT 804 by a conductor. DFE-PD 3520 may include one or more of the features described above with respect to DFE-PD 3420. In some examples, dielectric layer 3304, input and output regions 2202, 2204 and regions therebetween, and field plates 2820b, 2820c may be similar to those of DFE-PD 3420. However, in Fig.35 In the device of , the first substrate region 3302 below the dielectric layer 3304 (on which the DFE-PD 3520 is formed) may include a gate contact 2812 (HD region or ohmic contact) that allows the first substrate region 3302 to be electrically connected to the field plates 2820b, 2820c, thereby acting as a bottom field plate as part of the RESURF mechanism of the DFE-PD 3520. The L-BJT 804 may have a portion formed in each of the second and third substrate regions 204, 202, such as an oppositely doped well region, and extending across a boundary therebetween, wherein the second substrate region 204 forms a first vertical interface with the first substrate region 3302. The dielectric layer 218 disposed above the first interface electrically isolates the gate contact 2812 of the DFE-PD 3520 from the HD collector region 207 of the LV L-BJT 804. The second substrate region 204 may form a second vertical interface with the third substrate region 202. In some cases, the third substrate interface may be aligned with an edge of the base region of the LV L-BJT 804 .
[0413] The output region 2204 of the DFE-PD 3520 and the HD collector region of the LV L-BJT 804 can be electrically connected by a first conductive line 3000 fabricated in a metallization layer above the common substrate top surface. The field plates 2820b, 2820c, the gate contact 2812, and the RESURF layer 207 can be electrically connected by a second conductive line 3002 fabricated in a metallization layer above the common substrate top surface. The conductive lines 3002, 3000 can be patterned as part of the same or different metallization layers.
[0414] Advantageously, due to the electrical connection between the first substrate region 3302 and the field plates 2820 b , 2820 c , the first substrate region 3302 acts as a bottom field plate and contributes to the RESURF mechanism of the DFE-PD 3520 .
[0415] In some embodiments, the input and output regions 2202 , 2204 , the input and output drift regions 3312 , 2214 , and the second substrate region 204 may have the same polarity opposite to that of the first and third substrate regions 3302 , 202 , and the depletion control region 3220 .
[0416] In some cases, the first, second, and third substrate regions 3302 , 204 , 202 may be doped wells formed in a common substrate 290 .
[0417] In some cases, the combination of DFE-PD 3520 and L-BJT 804 (device within dashed box 3540 ) may electrically behave as a HV L-BJT, where input region 2202 serves as the effective collector region of the HV L-BJT.
[0418] As described above, the inventive aspect of being able to boost the low voltage BiCMOS platform to a suitable high voltage node involves creating high voltage DMOS and BJT devices using DFE-PD in series with lower voltage devices. In the above, the high voltage divider region is formed in a common substrate as the BiCMOS device. The inventors have discovered that by forming the DFE-PD in the back end of the line (BEOL) above the front end of the line (FEOL), further degrees of freedom in monolithic integration of different devices can be achieved.
[0419] According to various aspects, an integrated circuit (IC) device includes a metal oxide semiconductor (MOS) transistor and a bipolar junction transistor (e.g., L-BJT) and a lateral double diffused metal oxide semiconductor (LDMOS) formed in a common silicon substrate. The IC device also includes a DFE-PD or PD region formed above a main surface of the common semiconductor substrate and integrated in the back end (BEOL) of the IC device. In some embodiments, the high voltage divider region is formed of a wide bandgap semiconductor material. In some other embodiments, the high voltage divider region includes a main portion that includes a lightly doped semiconductor region. In some other embodiments, in operation, the high voltage divider region is configured to reduce a higher voltage at least relative to the MOS transistor.
[0420] According to various embodiments, the high voltage divider region is separated from the main surface of the common semiconductor substrate by at least one intermetallic dielectric layer. Similar to the above-mentioned in-substrate high voltage divider region, the high voltage divider region formed in the BEOL also includes a main portion, which includes a lightly (N or P) doped semiconductor region extending laterally between heavily doped regions. One of the heavily doped regions is connected to one or both of the BJT and LDMOS transistors, and the other heavily doped region is configured as a high voltage input and is configured to be biased at 5-400V. In addition, the high voltage divider region includes a conductive field plate formed thereon to serve as a reduced surface field (RESURF) plate. The high voltage divider region is formed of a wide bandgap semiconductor including one or more of SiC, GaN and Ga2O3.
[0421] Advantageously, by isolating the silicon substrate in which the low voltage devices are formed using a back-end dielectric layer, the high voltage nodes are effectively separated from the low voltage substrate. The inter-layer dielectric isolation can provide advantages such as vertical high voltage isolation, high voltage latch immunity, and reduction of parasitic capacitance.
[0422] In a similar manner as described above, various synergistic processing advantages can be achieved by manufacturing a metal oxide semiconductor (MOS) transistor and one or both of a bipolar junction transistor (BJT) and a lateral double diffused metal oxide semiconductor (LDMOS) transistor formed on a common silicon substrate. The details of the synergistic processing have been described above and will not be repeated here for the sake of brevity.
[0423] Fig.36 Schematically illustrates a cross-sectional view of another example high voltage device including a DFE-PD that is vertically separated and / or isolated from a substrate and electrically connected to an L-BJT via a wire. Example DFE-PD 3620 is electrically connected to LV L-BJT 804 via a wire and is vertically separated from a major surface (top surface) of substrate 290 on which LV L-BJT 804 is fabricated. In some implementations, DFE-PD 3620 may include the DFE-PD 3520 (described above) as described above. Fig.35 ) described above. DFE-PD 3620 is fabricated over a top surface of thick dielectric layer 3304 disposed on a top surface of substrate 290. In some cases, a vertical distance between a bottom surface of DFE-PD 3620 and a top surface of thick dielectric layer 3304 may be greater than 0.5 microns, greater than 1.0 microns, greater than 1.5 microns, greater than 2.0 microns, or a distance within a range defined by any of these values.
[0424] In some cases, DFE-PD 3620 may also include a bottom field plate 3600 disposed on a top surface of thick dielectric layer 3304. Bottom field plate 3600 may have a thickness (in the vertical direction) of 0.1 to 0.5 microns, or a thickness within a range defined by any of these values.
[0425] The output region 2204 of the DFE-PD 3620 and the HD collector region of the LV L-BJT 804 can be electrically connected by a first conductive line 3000 fabricated in a metallization layer over the top surface of the substrate 290. The field plates 2820b, 2820c, the bottom field plate 3600, and the RESURF layer 207 can be electrically connected by a second conductive line 3002 fabricated in a metallization layer over the common substrate top surface. In some examples, the bottom field plate 3600 can include an HD region that serves as an ohmic contact between the field plate 3600 and the second conductive line 3002.
[0426] The conductive lines 3002 , 3000 may be fabricated on the same or different metallization layers. The electrical connection between the bottom field plate 3600 and the field plates 2820 b , 2820 c results in an enhanced RESURF mechanism for the DFE-PD 3620 .
[0427] In some cases, the LV L-BJT 804 has portions of opposite polarity formed in each of the first and second regions 204, 202 of the substrate 290 and extending across a boundary therebetween. In some cases, the vertical interface between the two regions of the substrate 290 can be aligned with the vertical interface between the emitter and base regions of the L-BJT 804. The thick oxide layer 3304 can be disposed on the first region 204 of the substrate 290. In some cases, the first and second regions 204, 202 can be doped wells formed in the substrate 290, for example using thermal diffusion or ion implantation of dopants.
[0428] In some embodiments, the input and output regions 2202 , 2204 , the input and output drift regions 3312 , 2214 , the first substrate region 204 , the collector 207 and emitter 206 regions, and the HD region 3602 of the bottom field plate 3600 of the L-BJT 803 may have the same polarity, which is opposite to the polarity of the second substrate region 202 , the base well, and the depletion control region 3220 .
[0429] In some cases, the combination of DFE-PD 3620 and L-BJT 804 (device within dashed box 3640 ) may electrically behave as a HV L-BJT, where input region 2202 serves as the effective collector region of the HV L-BJT.
[0430] Fig.35 and 36The LV L-BJT 804 in may have the same arrangement of base well, base region, emitter region, RESURF layer and RESURF dielectric layer, and HD collector region as the L-BJT 804 described above with respect to FIG. 8 , except that the LD collector region is omitted.
[0431] Advantageously, physically separating the DFE-PD 3620 from the substrate on which the LV L-BJT 804 is fabricated provides more reliable electrical isolation between the DFE-PD 3620 and the LV L-BJT 804 (as well as other low voltage devices fabricated on the substrate 290). However, placing the bottom field plate 3600 on the thick dielectric layer 3304 can reduce substrate area that would otherwise be used to fabricate transistors and other active devices for performing logic, switching, and analog operations. In some embodiments, moving the DFE-PD further upward (e.g., increasing its vertical distance from the substrate) and separating the bottom field plate from the substrate can free up corresponding substrate area and allow more devices (e.g., BiCMOS devices) to be fabricated within a given substrate and connected to the DFE-PD for high voltage operation. In addition to freeing up surface area available for device fabrication, this approach can also replace polysilicon with various wide bandgap materials (e.g., bandgaps greater than 1.5 eV), such as compound semiconductors such as III-V and II-VI compound semiconductors, as structural materials for the DFE-PD to improve the quality of the junction formed between the depletion control region and the input / output drift region. In some embodiments, when the vertical distance between the DFE-PD and the substrate 290 is large enough, the DFE-PD structure can be formed using advanced material growth and deposition techniques (e.g., for processing large bandgap materials) that may not be available closer to the substrate. In addition, placing the DFE-PD at a higher level, such as above one or more metallization levels, facilitates electrical connections between the DFE-PD and multiple low-voltage devices through vias and lateral wire networks.
[0432] Fig.37A cross-sectional view of another example high voltage device is schematically shown, the device including a DFE-PD 3270 that is vertically separated and / or isolated from a substrate and electrically connected to one or more MOS transistors and BJTs. The example DFE-PD 3720 has top and bottom field plates 2820b, 3702 that are fabricated above one or more metallization layers (e.g., two metallization films) above a common substrate in which multiple MOS and / or bipolar devices are fabricated. For example, in the example shown, the DFE-PD 3720 is connected to the N-MOS 3702, the PNP L-BJT 3710, and the NPN L-BJT 3708 through the bottom field plate 3702 and parallel vertical electrical connections including through-holes between the output region 2204 and the drain and collector regions of the N-MOS 3702, the L-BJT 3710, and the L-BJT 3708.
[0433] Advantageously, the conductive lines extending laterally to connect the vias to the output region 2204 also serve as the bottom field plate 3702 of the DFE-PD 3720 .
[0434] In some embodiments, the vertical distance H between the input drift region 3312 of the DFE-PD 3720 and the top surface of the substrate on which the MOS and L-BJT transistors are fabricated can be greater than 0.15, greater than 0.3, greater than 0.5, greater than 1.0, or a distance within a range defined by any of these values.
[0435] In some cases, at least one region of DFE-PD 3720 may include a wide bandgap material, such as SiC, GaN, or Ga2O3.
[0436] In some cases, the thickness of the input region 2202, the input drift region 3312, and the depletion control region 3220 (of the DFE-PDs 3320, 3420, 3520, 3620) along the vertical direction (along the y-axis) can be substantially equal due to being formed from the same layer. In various embodiments, the thickness of the input region 2202, the input drift region 3312, and the depletion control region 3220 can be 0.15 to 0.5 microns. In some cases, the thickness of the dielectric layer 3304 (DFE-PDs 3320, 3420, 3520, 3620) can be 0.15 to 0.5 microns.
[0437] In various embodiments, the DFE PD 3320, 3420 ( Fig.34 )、3520( Fig.35 )、3620( Fig.36 ) and 3720( Fig.37)'s input region 2202, input drift region 3312, output drift region 2204 and output region 2204 may include a polysilicon layer or a wide bandgap material layer, which is disposed on the thick dielectric layer 3304 or on a portion of the dielectric cladding 214 above the thick dielectric layer 3304 or on the top surface of the substrate 290.
[0438] In various embodiments, the thickness of the polysilicon layer may be 0.1 to 0.15 microns, 0.1 to 0.3 microns, 0.1 to 1.0 microns, but less than 2.0 microns, or a thickness within a range defined by any of these values.
[0439] In some examples, the input region 2202, the input drift region 3312, the output drift region 224, and the output region 2204 of the DFE PD 3320, 3420, 3520, 3620, and 3720 may be formed in polysilicon or a wide bandgap semiconductor using a suitable doping technique, such as thermal diffusion, ion implantation, or in-situ doping. In some examples, the thickness of the input region 2202, the input drift region 3312, the output drift region 224, and the output region 2204 of the DFE PD 3320, 3420, 3520, 3620, and 3720 may extend vertically through the entire thickness of the polysilicon or wide bandgap semiconductor layer. In some cases, thermal diffusion or activation of dopants in the polysilicon layer may be achieved using a suitable technique, such as a rapid thermal annealing (RTA) process, which does not affect the wells and doped regions already formed in the substrate 290.
[0440] In some embodiments, Figure 31-36 The L-BJT in any of the configurations shown can be replaced with MOS transistors (e.g. Figure 7B and 7D In these embodiments, the output region 2204 of the corresponding DFE-PD (e.g., DFE-PD 3206, 3320, 3420, 3520, and 3620) can be electrically connected to the HD drain region 720 of the MOS transistor, wherein the HD drain region 720 and the output region 2204 have the same polarity. Therefore, the combination of the DFE-PD and the MOS transistor (the device within the dashed box 3240, 3340, 3440, 3540, and 3640) can electrically behave as a HV MOS or DMOS transistor, with the input region 2202 serving as the effective drain region of the DMOS.
[0441] In the above reference Fig.32In any of the embodiments described in -38, the thickness of the depletion control region 3220, the input drift region 3312, and the output drift region 3220 along the vertical direction may be substantially equal. In various embodiments, the thickness of the depletion control region 3220, the input drift region 3312, or the output drift region 3220 along the vertical direction may be 0.1 micrometers to 0.2 micrometers, 0.2 micrometers to 0.3 micrometers, 0.3 micrometers to 0.4 micrometers, 0.4 micrometers to 0.5 micrometers, or any value in the range formed by these values. In various embodiments, the common thickness of the depletion control region 3220, the input drift region 3312, or the output drift region 3220 along the vertical direction may be less than 2 micrometers.
[0442] In any of the embodiments described above with respect to FIGS. 23-37 , the majority carrier concentration or doping concentration in the inter-gate region 2215 and the depletion control region 3220 is 5, 10, 20, 30, 40, 50, 100% less than the majority carrier concentration or doping concentration in the input drift region 2212, 3212, 3312 and the output drift region 2214 (or the output drift region of the voltage divider 3420, 3520, 3620, 3720, or 3820) and / or the well 2828. 2 , 10 3 or larger value.
[0443] In any of the embodiments described above with respect to Figures 23-37, the majority carrier concentration or doping concentration in the inter-gate region 2215 and the depletion control region 3220 is at least 10 times smaller than the majority carrier concentration or doping concentration in the input drift region 2212, 3212, 3312 and the output drift region 2214 (or the output drift region of the voltage divider 3420, 3520, 3620, 3720 or 3820) and / or the well 2828.
[0444] Equivalent circuit model
[0445] Figures 38A-38B Without being bound by any theory, a circuit equivalent to the DFE-PD described above is shown. The secondary voltage source (V2, Vs) is a scaled function of the applied primary voltage (V1, Vp), and the primary current (I1) is determined by the secondary current (I2), where the current density for optimal performance is inversely proportional to the same scaled design parameters.
[0446] The voltage transformers shown are non-magnetic, have no active circuits, or will be applied with higher DC and AC voltage levels (primary, V p ) to a lower voltage (secondary, V s ) of a single component.
[0447] The ratio of the two voltage nodes is a design control parameter and can be represented by an integer N>1.
[0448]
[0449] The voltage transformer shown can be adjusted to target any voltage node, where the maximum planned voltage scaling converts 0-400V to 0-5V range. The output of the secondary terminal is directly connected to the low voltage (LV)-BiCMOS silicon substrate. The current i flowing through the LV secondary side of the voltage transformer s The operation of the LV circuit is controlled by the width W s and the height h represents the conduction area, which can be converted into current concentration:
[0450] i s =W s ×h×J s
[0451] No current flows through the higher voltage primary side i p , until the current flows through the lower voltage secondary side, where the current concentration ratio between the two sides is similarly scaled according to the design ratio N.
[0452]
[0453] According to the representative width W p and height h, the primary current concentration is related to the primary conduction area.
[0454] i p =W p ×h×J p
[0455] In order to optimally translate the performance achieved at a lower voltage node to a higher voltage node, the higher voltage must be able to support the same current conducted by the secondary:
[0456] i p =i s
[0457] This means that the conductive width of the higher voltage node must be N times the width of the lower voltage secondary:
[0458] W p ×h×J p =W s ×h×J s
[0459] This makes the area N of the primary high voltage node several times larger than the low voltage conduction area. The lower voltage area is limited by the MOS channel width or bipolar emitter terminal width. When the voltage transformer is conducting, it can behave like a resistive element. In the model shown, the combination of the LV BiCMOS components and the vertically stacked voltage transformers is drawn as a single device.
[0460] Some additional non-limiting examples of the above embodiments are provided below. These should not be construed as limiting the breadth of the disclosure in any way.
[0461] Example Embodiment I
[0462] 1. An integrated circuit (IC) device, comprising:
[0463] a metal oxide semiconductor (MOS) transistor including a gate stack formed on a channel region thereof; and
[0464] a bipolar junction transistor (BJT) comprising a layer stack formed on a collector region thereof,
[0465] The gate stack and the layer stack have one or more corresponding layers having common physical dimensions.
[0466] 2. An integrated circuit (IC) device comprising:
[0467] a metal oxide semiconductor (MOS) transistor including a gate stack formed on a channel region thereof; and
[0468] a bipolar junction transistor (BJT) comprising a layer stack formed on a collector region thereof,
[0469] The gate stack and the layer stack have corresponding spacer layer structures formed on sidewalls thereof and having common physical dimensions.
[0470] 3. An integrated circuit (IC) device comprising:
[0471] a metal oxide semiconductor (MOS) transistor including a gate stack formed on a channel region thereof; and
[0472] a bipolar junction transistor (BJT) comprising a layer stack formed on a collector region thereof,
[0473] The MOS transistor and the BJT have corresponding implant diffusion regions, and the implant diffusion regions have a common implant dopant distribution.
[0474] 4. The IC device of embodiment 1, wherein the gate stack and the layer stack have corresponding spacer layer structures formed on their sidewalls having common physical dimensions.
[0475] 5. The IC device of embodiment 1, wherein the MOS transistor and the BJT have corresponding implant diffusion regions having a common implant dopant profile.
[0476] 6. The IC device of embodiment 2, wherein the gate stack and the layer stack have one or more corresponding layers having common physical dimensions.
[0477] 7. The IC device of embodiment 2, wherein the MOS transistor and the BJT have corresponding implant diffusion regions having a common implant dopant distribution.
[0478] 8. The IC device of embodiment 3, wherein the gate stack and the layer stack have corresponding spacer layer structures formed on their sidewalls having common physical dimensions.
[0479] 9. The IC device of embodiment 8, wherein the corresponding implanted diffusion region includes a source region and a drain region of a MOS transistor and one of an emitter region or a collector region of a BJT.
[0480] 10. The IC device of embodiment 1, wherein the common physical dimensions result from one or more of co-oxidation, co-deposition, co-etching, and co-patterning of the gate stack and the layer stack.
[0481] 11. The IC device of embodiment 2, wherein the common physical dimensions result from one or more of co-deposition, co-etching, and co-patterning of corresponding spacer layer structures.
[0482] 12. The IC device of embodiment 3, wherein the common implant dopant profile is caused by one or more of co-implantation and co-dopant activation of corresponding implant diffusion regions.
[0483] 13. The IC device of any of the above embodiments, wherein the MOS transistor and the BJT are monolithically integrated on a common substrate.
[0484] 14. An IC device according to any one of the above embodiments, wherein the BJT is a lateral BJT having an emitter region, a base region in a base well, and a collector region in a common well, wherein the emitter region, the base region, and the collector region are arranged in a lateral direction to support a lateral bipolar current, and wherein the lateral direction is parallel to a main surface of a common substrate on which the BJT and the MOS transistor are manufactured.
[0485] 15. The IC device of any one of the preceding embodiments, wherein a base length of the BJT extends laterally from a first end at a vertical interface between the base region and the emitter region to a second end below an edge of the layer stack.
[0486] 16. An IC device according to any of the above embodiments, wherein the layer stack of the BJT has a first RESURF dielectric layer, which is co-deposited or co-oxidized with the first gate dielectric layer of the MOS transistor and has the same thickness as the first gate dielectric layer of the MOS transistor.
[0487] 17. The IC device of any one of the preceding embodiments, wherein the layer stack of the BJT has a RESURF layer on the first RESURF dielectric layer, the RESURF layer being co-deposited with a gate electrode of the first gate dielectric layer of the MOS transistor and having the same thickness.
[0488] 18. The IC device of any one of the above embodiments, wherein at least a portion of the RESURF layer is disposed on a second RESURF dielectric layer, the second RESURF dielectric layer being co-deposited with a second gate dielectric layer of the MOS transistor and having the same thickness as the gate layer.
[0489] 19. An IC device according to any of the above embodiments, wherein the thickness of the second RESURF dielectric layer and the second gate dielectric layer along the vertical direction is respectively greater than the thickness of the first RESURF dielectric layer and the first gate dielectric layer, wherein the vertical direction is perpendicular to the main surface of the common substrate on which the BJT and MOS transistors are manufactured.
[0490] 20. The IC device of any one of the previous embodiments, wherein the second RESURF dielectric layer and the second gate dielectric layer include thermally grown oxide layers.
[0491] 21. An IC device according to any one of the above embodiments, wherein the lateral dimension of the layer stack in the first direction defines the drift region or collector length of the BJT, and wherein the longitudinal dimension of the gate stack in the second direction defines the channel length of the MOS transistor in the first direction.
[0492] 22. The IC device of any one of the preceding embodiments, wherein the first direction is parallel to the second direction.
[0493] 23. The IC device according to any one of the above embodiments, wherein a length of the layer stack in a lateral direction is...
Claims
1. An integrated circuit (IC) device configured to reduce a voltage between an input and an output, the IC device comprising: An isolation substrate region is formed in the semiconductor substrate and is electrically isolated from the semiconductor substrate in the vertical and lateral directions. A plurality of alternating doping regions are formed in the isolation substrate region. The alternating doping regions are laterally arranged in the lateral direction and are alternately doped with opposite types of dopants. The alternating doping regions include: an input drift region and an output drift region, each doped with a first type of dopant, wherein the input drift region is connected to the input and the output drift region is connected to the output, an inter-gate region and a substrate region isolating the substrate region, each doped with a second type of dopant, wherein the inter-gate region is laterally between the input and output drift regions, and a heavily doped first gate region formed in the inter-gate region; The input drift region is elongated to have a first lateral length along the lateral direction, the first lateral length being at least twice longer than a second lateral length of the inter-gate region. 2 . The IC device of claim 1 , further comprising a dielectric layer covering the input drift region, and at least one conductive field plate extending laterally above the dielectric layer. 3 . The IC device according to claim 1 , wherein the second type dopant concentration of the inter-gate region is lower than the concentration of the substrate region of the isolation substrate region outside the input drift region, the output drift region and the alternating doping region. The IC device of claim 2 , wherein at least one of the conductive field plates is electrically connected to the first gate region.
5. The IC device according to any one of claims 1 to 4 further includes a heavily doped gate contact region formed at a surface of a substrate region of the isolation substrate region, the heavily doped gate contact region being configured to have the same polarity as the heavily doped first gate region, wherein the substrate region is electrically connected to the highly doped first gate and the heavily impurity gate contact region to serve as a second gate region.
6. The IC device of any one of claims 1 to 5, wherein the IC device is configured to provide an output voltage greater than 1 microvolt to the output in response to receiving an input voltage greater than 0.5 volts from the input.
7. The IC device of any one of claims 1 to 6, wherein the voltage reduction factor between the input and the output is at least 2.
8. An integrated circuit (IC) device configured to reduce a voltage between an input and an output, the IC device comprising: An isolation substrate region is formed in the semiconductor substrate and is electrically isolated from the semiconductor substrate in the vertical and lateral directions. A plurality of alternating doping regions are formed in the isolation substrate region. The alternating doping regions are laterally arranged in the lateral direction and are alternately doped with opposite types of dopants. The alternating doping regions include: an input drift region and an output drift region, each doped with a first type of dopant, wherein the input drift region is connected to the input and the output drift region is connected to the output, and an inter-gate region and a substrate region isolating the substrate region, each doped with a second type of dopant, wherein the inter-gate region is laterally between the input and output drift regions, wherein the dopant concentration of the inter-gate region is lower than the doping concentrations of the input drift region, the output drift region and the substrate region, and A heavily doped first gate region is formed in the inter-gate region. 9 . The IC device of claim 8 , further comprising a dielectric layer covering the input drift region, and one or more conductive field plates extending laterally above the dielectric layer. 10 . The IC device of claim 8 , wherein the input drift region has a first lateral length along the lateral direction, the first lateral length being at least twice longer than a second lateral length of the second doping region. 11 . The IC device of claim 9 , wherein at least one of the conductive field plates is electrically connected to the first gate region.
12. The IC device according to any one of claims 8 to 11 further includes a heavily doped gate contact region formed at a surface of a substrate region of the isolation substrate region, the heavily doped gate contact region being configured to have the same polarity as the heavily doped first gate region, wherein the substrate region is electrically connected to the highly doped first gate and the heavily impurity gate contact region to serve as a second gate region.
13. The IC device of any one of claims 8 to 12, wherein the IC device is configured to provide an output voltage greater than 1 microvolt to the output in response to receiving an input voltage greater than 0.5 volts from the input.
14. An integrated circuit (IC) device configured to reduce a voltage between an input and an output, the IC device comprising: An isolation substrate region is formed in the semiconductor substrate and is electrically isolated from the semiconductor substrate in the vertical and lateral directions. A plurality of alternating doping regions are formed in the isolation substrate region. The alternating doping regions are laterally arranged in the lateral direction and are alternately doped with opposite types of dopants. The alternating doping regions include: an input drift region and an output drift region, each doped with a first type of dopant, wherein the input drift region is connected to the input and the output drift region is connected to the output, an inter-gate region and a substrate region isolating the substrate region, each doped with a second type of dopant, wherein the inter-gate region is laterally between the input and output drift regions, a dielectric layer covering the input drift region, and at least one conductive field plate extending laterally above the dielectric layer, and A heavily doped (HD) first gate region is formed in the inter-gate region. 15 . The IC device of claim 14 , wherein a concentration of the second type dopant in the inter-gate region is lower than a concentration in a remaining portion of the isolation substrate region outside the input drift region, the output drift region, and the alternately doped region. 16 . The IC device of claim 14 , wherein the input drift region has a first lateral length along the lateral direction, the first lateral length being at least twice longer than a second lateral length of the second doped region.
17. The IC device of any one of claims 14 to 16, wherein at least one of the conductive field plates is electrically connected to the first gate region.
18. The IC device according to any one of claims 14 to 17 further includes a heavily doped gate contact region formed at a surface of a substrate region of the isolation substrate region, the heavily doped gate contact region being configured to have the same polarity as the heavily doped first gate region, wherein the substrate region is electrically connected to the highly doped first gate and the heavily impurity gate contact region to serve as a second gate region.
19. The IC device of any one of claims 14 to 18, wherein the IC device is configured to provide an output voltage greater than 1 microvolt to the output in response to receiving an input voltage greater than 0.5 volts from the input.
20. The IC device of any one of claims 14 to 19, wherein the isolated substrate region is electrically isolated from other substrate regions by at least one of a dielectric-filled trench and a buried dielectric layer.
