USRE48965E1

USRE48965E1 - Method and apparatus improving gate oxide reliability by controlling accumulated charge

Info

Publication number: USRE48965E1
Application number: US16/710,998
Authority: US
Inventors: Michael A. Stuber; Christopher N. Brindle; Dylan J. Kelly; Clint L. Kemerling; George P. Imthurn; Robert B. Welstand; Mark L. Burgener; Alexander Dribinsky; Tae Youn Kim
Original assignee: PSemi Corp
Current assignee: PSemi Corp
Priority date: 2005-07-11
Filing date: 2019-12-11
Publication date: 2022-03-08
Anticipated expiration: 2026-07-10
Also published as: US11201245B2; US20240266441A1; US20220181497A1; US20200335633A1; US11901459B2

Abstract

A method and apparatus are disclosed for use in improving the gate oxide reliability of semiconductor-on-insulator (SOI) metal-oxide-silicon field effect transistor (MOSFET) devices using accumulated charge control (ACC) techniques. The method and apparatus are adapted to remove, reduce, or otherwise control accumulated charge in SOI MOSFETs, thereby yielding improvements in FET performance characteristics. In one embodiment, a circuit comprises a MOSFET, operating in an accumulated charge regime, and means for controlling the accumulated charge, operatively coupled to the SOI MOSFET. A first determination is made of the effects of an uncontrolled accumulated charge on time dependent dielectric breakdown (TDDB) of the gate oxide of the SOI MOSFET. A second determination is made of the effects of a controlled accumulated charge on TDDB of the gate oxide of the SOI MOSFET. The SOI MOSFET is adapted to have a selected average time-to-breakdown, responsive to the first and second determinations, and the circuit is operated using techniques for accumulated charge control operatively coupled to the SOI MOSFET. In one embodiment, the accumulated charge control techniques include using an accumulated charge sink operatively coupled to the SOI MOSFET body.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS Claims of Priority

This application is a reissue of U.S. application Ser. No. 13/028,144, filed Feb. 15, 2011, now U.S. Pat. No. 8,954,902, which is a divisional application of U.S. application Ser. No. 11/520,912, “Method and Apparatus Improving Gate Oxide Reliability by Controlling Accumulated Charge”, filed Sep. 14, 2006, issuing Feb. 15, 2011 as U.S. Pat. No. 7,890,891, which is a Continuation-in-Part (CIP) of commonly assigned U.S. patent application Ser. No. 11/484,370, filed Jul. 10, 2006, entitled “Method and Apparatus for use in Improving Linearity of MOSFETs using an Accumulated Charge Sink”, which claims the benefit under 35 U.S.C. §119 (e) of U.S. Provisional Application No. 60/698,523, filed Jul. 11, 2005, entitled “Method and Apparatus for use in Improving Linearity of MOSFETs using an Accumulated Charge Sink”; and application Ser. No. 11/520,912, also claims the benefit under 35 U.S.C. §119 (e) of U.S. Provisional Application No. 60/718,260, filed Sep. 15, 2005, entitled “Method and Apparatus Improving Gate Oxide Reliability by Controlling Accumulated Charge”; and the contents of application Ser. No. 11/484,370 and 11/520,912, and provisional application Nos. 60/698,523 and 60/718,260, are all incorporated by reference herein in their entirety.

BACKGROUND

1. Field

The present invention relates to metal-oxide-semiconductor (MOS) field effect transistors (FETs), and particularly to MOSFETs fabricated on Semiconductor-On-Insulator (“SOI”) and Silicon-On-Sapphire (“SOS”) substrates. In one embodiment, an SOI (or SOS) MOSFET is adapted to control accumulated charge and thereby improve linearity of circuit elements. In another embodiment, according to the present CIP, an SOI (or SOS) MOSFET is adapted to control accumulated charge and thereby improve gate oxide reliability.

2. Description of Related Art

Although the disclosed method and apparatus for use in improving the linearity of MOSFETs are described herein as applicable for use in SOI MOSFETs, it will be appreciated by those skilled in the electronic device design arts that the present teachings are equally applicable for use in SOS MOSFETs. The present teachings also apply to other semiconductor-on-insulator systems, wherein the silicon is replaced by another semiconductor such as silicon-germanium (SiGe). It will also be appreciated by those skilled in the electronic design arts that the present disclosed method and apparatus also apply to virtually any insulating gate technology, and to integrated circuits having a floating body. As those skilled in the art will appreciate, technologies are constantly being developed for achieving “floating body” implementations. For example, the inventors are aware of circuits implemented in bulk silicon wherein circuit implementations are used to “float” the body of the device. The disclosure contemplates embodiments of the disclosed method and apparatus implemented in any of the developing floating body implementations. Therefore, references to and exemplary descriptions of SOI MOSFETs herein are not to be construed as limiting the applicability of the present teachings to SOI MOSFETs only. Rather, as described below in more detail, the disclosed method and apparatus find utility in MOSFETs implemented in a plurality of device technologies, including SOS.

As is well known, a MOSFET employs a gate-modulated conductive channel of n-type or p-type conductivity, and is accordingly referred to as an “NMOSFET” or “PMOSFET”, respectively. FIG. 1 shows a cross-sectional view of an exemplary prior art SOI NMOSFET 100. As shown in FIG. 1, the prior art SOI NMOSFET 100 includes an insulating substrate 118 that may comprise a buried oxide layer, sapphire, or other insulating material. A source 112 and drain 116 of the NMOSFET 100 comprise N+ regions (i.e., regions that are heavily doped with an “n-type” dopant material) produced by ion implantation into a silicon layer positioned above the insulating substrate 118. (The source and drain of PMOSFETs comprise P+ regions (i.e., regions heavily doped with “p-type” dopant material)). The body 114 comprises a P− region (i.e., a region that is lightly doped with a “p-type” dopant), produced by ion implantation, or by dopants already present in the silicon layer when it is formed on the insulating substrate 118. As shown in FIG. 1, the NMOSFET 100 also includes a gate oxide 110 positioned over the body 114. The gate oxide 110 typically comprises a thin layer of an insulating dielectric material such as SiO₂. The gate oxide 110 electrically insulates the body 114 from a gate 108 positioned over the gate oxide 110. The gate 108 comprises a layer of metal or, more typically, polysilicon.

A source terminal 102 is operatively coupled to the source 112 so that a source bias voltage “Vs” may be applied to the source 112. A drain terminal 106 is operatively coupled to the drain 116 so that a drain bias voltage “Vd” may be applied to the drain 116. A gate terminal 104 is operatively coupled to the gate 108 so that a gate bias voltage “Vg” may be applied to the gate 108.

As is well known, when a voltage is applied between the gate and source terminals of a MOSFET, a generated electric field penetrates through the gate oxide to the transistor body. For an enhancement mode device, a positive gate bias creates a channel in the channel region of the MOSFET body through which current passes between the source and drain. For a depletion mode device, a channel is present for a zero gate bias. Varying the voltage applied to the gate modulates the conductivity of the channel and thereby controls the current flow between the source and drain.

For an enhancement mode MOSFET, for example, the gate bias creates a so-called “inversion channel” in a channel region of the body 114 under the gate oxide 110. The inversion channel comprises carriers having the same polarity (e.g., “P” polarity (i.e., hole carriers), or “N” polarity (i.e., electron carriers) carriers) as the polarity of the source and drain carriers, and it thereby provides a conduit (i.e., channel) through which current passes between the source and the drain. For example, as shown in the SOI NMOSFET 100 of FIG. 1, when a sufficiently positive voltage is applied between the gate 108 and the source 112 (i.e. a positive gate bias exceeding a threshold voltage V_th), an inversion channel is formed in the channel region of the body 114. As noted above, the polarity of carriers in the inversion channel is identical to the polarity of carriers in the source and drain. In this example, because the source and drain comprise “n-type” dopant material and therefore have N polarity carriers, the carriers in the channel comprise N polarity carriers. Similarly, because the source and drain comprise “p-type” dopant material in PMOSFETs, the carriers in the channel of turned on (i.e., conducting) PMOSFETs comprise P polarity carriers.

Depletion mode MOSFETs operate similarly to enhancement mode MOSFETs, however, depletion mode MOSFETs are doped so that a conducting channel exists even without a voltage being applied to the gate. When a voltage of appropriate polarity is applied to the gate the channel is depleted. This, in turn, reduces the current flow through the depletion mode device. In essence, the depletion mode device is analogous to a “normally closed” switch, while the enhancement mode device is analogous to a “normally open” switch. Both enhancement and depletion mode MOSFETs have a gate voltage threshold, V_th, at which the MOSFET changes from an off-state (non-conducting) to an on-state (conducting).

No matter what mode of operation an SOI MOSFET employs (i.e., whether enhancement or depletion mode), when the MOSFET is operated in an off-state (i.e., the gate voltage does not exceed V_th), and when a sufficient nonzero gate bias voltage is applied with respect to the source and drain, an “accumulated charge” may occur under the gate. The “accumulated charge”, as defined in more detail below and used throughout the present application, is similar to the “accumulation charge” described in the prior art literature in reference to MOS capacitors. However, the prior art references describe “accumulation charge” as referring only to bias-induced charge existing under a MOS capacitor oxide, wherein the accumulation charge is of the same polarity as the majority carriers of the semiconductor material under the capacitor oxide. In contrast, and as described below in more detail, “accumulated charge” is used herein to refer to gate-bias induced carriers that may accumulate in the body of an off-state MOSFET, even if the majority carriers in the body do not have the same polarity as the accumulated charge. This situation may occur, for example, in an off-state depletion mode NMOSFET, wherein the accumulated charge may comprise holes (i.e., having P polarity) even though the body doping is N− rather than P−.

For example, as shown in FIG. 1, when the SOI NMOSFET 100 is biased to operate in an off-state, and when a sufficient nonzero voltage is applied to the gate 108, an accumulated charge 120 may accumulate in the body 114 underneath and proximate the gate oxide 110. The operating state of the SOI NMOSFET 100 shown in FIG. 1 is referred to herein as an “accumulated charge regime” of the MOSFET. The accumulated charge regime is defined in more detail below. The causes and effects of the accumulated charge in SOI MOSFETs are now described in more detail.

As is well known, electron-hole pair carriers may be generated in MOSFET bodies as a result of several mechanisms (e.g., thermal, optical, and band-to-band tunneling electron-hole pair generation processes). When electron-hole pair carriers are generated within an NMOSFET body, for example, and when the NMOSFET is biased in an off-state condition, electrons may be separated from their hole counterparts and pulled into both the source and drain. Over a period of time, assuming the NMOSFET continues to be biased in the offstate, the holes (resulting from the separated electron-hole pairs) may accumulate under the gate oxide (i.e., forming an “accumulated charge”) underneath and proximate the gate oxide. A similar process (with the behavior of electrons and holes reversed) occurs in similarly biased PMOSFET devices. This phenomenon is now described with reference to the SOI NMOSFET 100 of FIG. 1.

When the SOI NMOSFET 100 is operated with gate, source and drain bias voltages that deplete the channel carriers in the body 114 (i.e., the NMOSFET 100 is in the off-state), holes may accumulate underneath and proximate the gate oxide 110. For example, if the source bias voltage Vs and the drain bias voltage Vd are both zero (e.g., connected to a ground contact, not shown), and the gate bias voltage Vg comprises a sufficiently negative voltage with respect to ground and with respect to V_th, holes present in the body 114 become attracted to the channel region proximate the gate oxide 110. Over a period of time, unless removed or otherwise controlled, the holes accumulate underneath the gate oxide 110 and result in the accumulated charge 120 shown in FIG. 1. The accumulated charge 120 is therefore shown as positive “+” hole carriers in FIG. 1. In the example given, Vg is negative with respect to Vs and Vd, so

electric field regions

122 and 124 may also be present.

Accumulated Charge Regime Defined

The accumulated charge is opposite in polarity to the polarity of carriers in the channel. Because, as described above, the polarity of carriers in the channel is identical to the polarity of carriers in the source and drain, the polarity of the accumulated charge 120 is also opposite to the polarity of carriers in the source and drain. For example, under the operating conditions described above, holes (having “P” polarity) accumulate in off-state NMOSFETs, and electrons (having “N” polarity) accumulate in off-state PMOSFETs. Therefore, a MOSFET device is defined herein as operating within the “accumulated charge regime” when the MOSFET is biased to operate in an off-state, and when carriers having opposite polarity to the channel carriers are present in the channel region. Stated in other terms, a MOSFET is defined as operating within the accumulated charge regime when the MOSFET is biased to operate in an off-state, and when carriers are present in the channel region having a polarity that is opposite the polarity of the source and drain carriers.

For example, and referring again to FIG. 1, the accumulated charge 120 comprises hole carriers having P or “+” polarity. In contrast, the carriers in the source, drain, and channel (i.e., when the FET is in the on-state) comprise electron carriers having N or “−” polarity. The SOI NMOSFET 100 is therefore shown in FIG. 1 as operating in the accumulated charge regime. It is biased to operate in an off-state, and an accumulated charge 120 is present in the channel region. The accumulated charge 120 is opposite in polarity (P) to the polarity of the channel, source and drain carriers (N).

In another example, wherein the SOI NMOSFET 100 comprises a depletion mode device, V_this negative by definition. According to this example, the body 114 comprises an N− region (as contrasted with the P− region shown in FIG. 1). The source and drain comprise N+ regions similar to those shown in the enhancement mode MOSFET 100 of FIG. 1. For Vs and Vd both at zero volts, when a gate bias Vg is applied that is sufficiently negative relative to V_th(for example, a Vg that is more negative than approximately −1 V relative to V_th), the depletion mode NMOSFET is biased into an off-state. If biased in the off-state for a sufficiently long period of time, holes may accumulate under the gate oxide and thereby comprise the accumulated charge 120 shown in FIG. 1.

In other examples, Vs and Vd may comprise nonzero bias voltages. In some embodiments, Vg must be sufficiently negative to both Vs and Vd (in order for Vg to be sufficiently negative to V_th, for example) in order to bias the NMOSFET in the off-state. Those skilled in the MOSFET device design arts shall recognize that a wide variety of bias voltages may be used to practice the present teachings. As described below in more detail, the present disclosed method and apparatus contemplates use in any SOT MOSFET device biased to operate in the accumulated charge regime.

SOI and SOS MOSFETs are often used in applications in which operation within the accumulated charge regime adversely affects MOSFET performance. As described below in more detail, unless the accumulated charge is removed or otherwise controlled, it detrimentally affects performance of SOI MOSFETs under certain operating conditions. One exemplary application, described below in more detail with reference to the circuits shown in FIGS. 2B and 5A, is the use of SOI MOSFETs in the implementation of radio frequency (RF) switching circuits. As described below with reference to FIGS. 2B and 5A in more detail, the inventors have discovered that unless the accumulated charge is removed or otherwise controlled, under some operating conditions, the accumulated charge adversely affects the linearity of the SOT MOSFET and thereby increases harmonic distortion and intermodulation distortion (IMD) caused by the MOSFET when used in the implementation of certain circuits. In addition, as described below in more detail, the inventors have discovered that removal or control of the accumulated charge improves the drain-to-source breakdown voltage (i.e., the “BVDSS”) characteristics of the SOI MOSFETs.

Therefore, it is desirable to provide techniques for adapting and improving SOI (and SOS) MOSFETs, and circuits implemented with the improved SOI MOSFETs, in order to remove or otherwise control the accumulated charge, and thereby significantly improve SOI MOSFET performance. It is desirable to provide methods and apparatus for use in improving the linearity characteristics in SOI MOSFETs. The improved MOSFETs should have improved linearity, harmonic distortion, intermodulation distortion, and BVDSS characteristics as compared with prior art MOSFETs, and thereby improve the performance of circuits implemented with the improved MOSFETs. The present teachings provide such novel methods and apparatus.

Gate Oxide Reliability and the Accumulated Charge Regime

The gate oxide is a critical component of a MOSFET. In many applications, including RF switch implementation, it is desirable to make the gate oxide as thin as possible. In RF circuit applications, thinner gate oxide results in higher on-currents and lower insertion losses for RF signals. However, if the gate oxide is too thin, the oxide will break down when a gate voltage is applied. When an electric field is applied to a gate oxide, there is typically a significant time interval before the gate oxide fails. The time required for a gate oxide to fail is a function of the applied electric field and temperature. This phenomenon is known as Time Dependent Dielectric Breakdown (TDDB). As a rough rule of thumb, at room temperature the electric field in a gate oxide should not exceed approximately 5 MV/cm for a desired lifetime or time-to-breakdown of ten years.

TDDB in gate oxides has been investigated extensively. One exemplary reference is an article entitled “A Unified Gate Oxide Reliability Model,” C. Hu and Q. Lu, 37th International Reliability Physics Symposium, San Diego, Calif. 1999. This paper discusses two major mechanisms for TDDB which occur under different stress conditions related to the strength of the applied electric field.

Another exemplary reference is a technical paper entitled “Low Electric Field Breakdown of Thin SiO2 Films Under Static and Dynamic Stress,” J. S. Suehle and P. Chaparala, IEEE Transactions on Electron Devices, Vol. 44, No. 5, May 1997. This reference reports an increase in gate oxide lifetime under bipolar pulsed stress (positive and negative voltage pulses) relative to unipolar (DC) stress. This effect occurs only at very large fields, and is attributed to relaxation of hole trapping occurring in the gate oxide. This phenomenon is unrelated to the improvements in gate oxide reliability that can be obtained by controlling accumulated charge in SOI MOSFETs, as described in more detail herein.

Still yet another reference relating to TDDB is a technical paper entitled “Reliability Issues for Silicon-on-insulator,” R. Bolam, et al., Electron Devices Meeting 2000, IEDM Technical Digest, December 2000. The authors report that there is no significant difference for TDDB failure in bulk Si devices and SOI devices fabricated in accordance with current art. The paper by C. Hu and Q. Lu, the reference by Suehle and Chaparala, and the reference by R. Bolam, cited above (referred to herein as the “TDDB references”), are hereby fully incorporated by reference herein, as though set forth in full for their teachings on the reliability of SiO₂when used as a gate dielectric.

The TDDB references cited above indicate that TDDB lifetime, at a given temperature, is dependent on the electric field in the gate oxide. When charge carriers are present beneath the gate oxide (e.g., when the MOSFET is in an on-state), the electric field in the oxide is approximately equal to the gate-to-source voltage divided by the gate oxide thickness. However, in accordance with teachings presented herein, persons skilled in the arts of electronic devices will appreciate that this is not necessarily the case for an SOI MOSFET operated in an off-state in the accumulated charge regime. In this case, the electric field that stresses the gate oxide is also affected by the presence of an accumulated charge under the gate. In particular, the inventors have discovered that removing or otherwise controlling the accumulated charge can significantly reduce the electric field that stresses the gate oxide and thereby improve the gate oxide reliability. Therefore, it is desirable to provide techniques for adapting and operating SOI MOSFET devices and circuits in order to control the accumulated charge and thereby significantly improve gate oxide reliability. The present teachings provide such novel techniques for adapting and operating SOI MOSFET devices.

SUMMARY

Apparatuses and methods are provided to control accumulated charge in SOI MOSFETs, thereby improving nonlinear responses and harmonic and intermodulation distortion effects in the operation of the SOI MOSFETs.

In one embodiment, a circuit having at least one SOI MOSFET is configured to operate in an accumulated charge regime. An accumulated charge sink (ACS), operatively coupled to the body of the SOI MOSFET, receives accumulated charge generated in the body, thereby reducing the non-linearity of the net source-drain capacitance of the SOI MOSFET.

In one embodiment, the ACS comprises a high impedance connection to the MOSFET body, with an exemplary impedance greater than 10⁶ohm.

Embodiments Relating to Methods and Apparatuses for Improving Gate Oxide Reliability in Accordance with the Present CIP

Circuits with SOI MOSFETs are adapted to control accumulated charge in the SOI MOSFETs, thereby improving gate oxide reliability of the SOI MOSFETs.

In one embodiment, a circuit comprises an SOI MOSFET, operating in the accumulated charge regime, and means for accumulated charge control (ACC), operatively coupled to the SOI MOSFET. The SOI MOSFET is adapted to have a maximum electric field E_tbin the gate oxide of the SOI MOSFET, where E_tbcorresponds to a desired lifetime or average time-to-breakdown for the gate oxide. The SOI MOSFET is adapted responsive to a first determination of the maximum electric field E_ox1in the gate oxide with an uncontrolled accumulated charge proximate to the gate oxide, and further responsive to second determination of the maximum electric field E_ox2in the gate oxide with a controlled accumulated charge proximate to the gate oxide. Determinations of E_tb, E_ox1, and E_ox2may be performed using well known techniques such as TDDB measurements and simulations of the SOI MOSFET operation.

According to another embodiment, a method for improving gate oxide reliability of an SOI MOSFET, operating in the accumulated charge regime, begins at a STEP (a). At the STEP (a), a first maximum electric field E_ox1in the gate oxide of the SOI MOSFET is determined with an uncontrolled accumulated charge proximate to the gate oxide. Proceeding to a STEP (b), a second maximum electric field E_ox2in the gate oxide of the SOI MOSFET is determined with a controlled accumulated charge proximate to the gate oxide. At a STEP (c), the SOI MOSFET is implemented in a circuit, wherein the SOI MOSFET is adapted to have a maximum electric field E_tbin the gate oxide. The maximum electric field E_tbcorresponds to a desired lifetime or time-to-breakdown for the gate oxide. The SOI MOSFET is adapted responsive to the determinations of the STEPS (a) and (b). At a final STEP (d), the circuit is operated using a means for ACC operatively coupled to the SOI MOSFET.

In another embodiment, the SOI MOSFET is adapted by implementing a second thickness T_ox2of the gate oxide, wherein T_ox2is less than a first thickness T_ox1of the gate oxide. The first thickness T_ox1corresponds to the SOI MOSFET having the maximum electric field E_ox1in the gate oxide less than or equal to E_tbwhen the SOI MOSFET is operated without the means for ACC operatively coupled to the SOI MOSFET. Implementation of the means for ACC enables the use of the lesser second thickness T_ox2, without allowing the maximum electric field E_ox2in the gate oxide to exceed E_tb.

According to another embodiment, the SOI MOSFET is adapted to enable a specified performance for the SOI MOSFET using the second thickness T_ox2and a second body width W2 of the SOI MOSFET (the terms “gate width,” “body width” and “transistor width” are used equivalently and interchangeably herein), wherein the second body width W2 is less than a first body width W1 of the SOI MOSFET. The first body width W1 corresponds to the SOI MOSFET being adapted to enable the specified performance using the first thickness T_ox1. Using the lesser second gate oxide thickness T_ox2provides improvements in the performance of the SOI MOSFET, such as reduced insertion loss, improved on-state conductance, and improved transconductance. This allows the body width of the SOI MOSFET to be reduced while maintaining the specified performance. Advantages of reduced body width include smaller die size for lower manufacturing cost and reduced parasitic capacitance, and better performance such as improved linearity.

In a further embodiment, the SOI MOSFET is adapted by implementing a second limiting gate bias voltage V_g2applied to a gate of the SOI MOSFET. V_g2has an absolute value greater than the absolute value of a first limiting gate bias voltage V_g1applied to the gate of the SOI MOSFET. V_g1corresponds to the SOI MOSFET having the maximum electric field E_ox1in the gate oxide less than or equal to E_tbwhen the SOI MOSFET is operated without the means for ACC operatively coupled to the SOI MOSFET. Implementation of the means for ACC enables the magnitude of the off-state limiting gate bias V_g2to be increased over V_g1without having the maximum electric field E_ox2in the gate oxide exceed E_tb. More generally, using ACC allows the use of thinner gate oxide, larger gate bias magnitude, or a selected combination thereof.

In another embodiment, the means for ACC may comprise an Accumulated Charge Sink (ACS), having a resistance greater than 10⁶ohm, operatively coupled to a body region of the SOI MOSFET to receive and remove the accumulated charge.

In another exemplary embodiment, the means for ACC may comprise an ACS, having a resistance not greater than 10⁶ohm, operatively coupled to a body region of the SOI MOSFET to receive and remove the accumulated charge.

In yet another exemplary embodiment, the means for ACC may comprise a control circuit operatively coupled to a gate of the SOI MOSFET. The control circuit is adapted to apply a voltage pulse to the gate to switch the SOI MOSFET from the off-state to an on-state for a selected interval, thereby dissipating the accumulated charge proximate to the gate oxide.

In yet another exemplary embodiment, the means for ACC may comprise a resistor electrically connected to a gate of the SOI MOSFET sufficient to prevent attenuation of an induced RF voltage on the gate, and wherein an RF signal voltage is applied to a drain of the SOI MOSFET that generates the induced RF voltage, thereby reducing the accumulated charge.

According to other exemplary embodiments, the SOI MOSFET may be an NMOSFET or a PMOSFET. The SOI MOSFET may be fabricated using silicon-on-sapphire, or other SOI technologies, such as Separation by Implantation of Oxygen (SIMOX)), silicon bonded to insulator, etc. The exemplary circuit including the SOI MOSFET may be an RF switching circuit, an RF mixer, a power amplifier, a level shifting circuit, a negative voltage generator, an oscillator, a DC-DC converter or other circuit using SOI MOSFETs.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a cross-sectional view of an exemplary prior art SOI NMOSFET.

FIG. 2A is a simplified schematic of an electrical model showing the off-state impedance characteristics of the exemplary prior art SOI NMOSFET of FIG. 1.

FIG. 2B is a schematic of an exemplary simplified RF switching circuit implemented using prior art SOI MOSFETs such as the prior art SOI NMOSFET of FIG. 1.

FIGS. 3A and 3B are simplified schematic diagrams of a top view of an improved SOI NMOSFET adapted to control accumulated charge in accordance with the present teachings.

FIG. 3C is a cross-sectional perspective schematic of an improved SOI NMOSFET adapted to control accumulated charge showing gate, source, drain and accumulated charge sink (ACS) terminals.

FIG. 3D is a simplified top view schematic of an improved SOI NMOSFET adapted to control accumulated charge having an accumulated charge sink (ACS) electrically coupled to a P+ region.

FIG. 3E is a simplified top view schematic of an improved SOI NMOSFET adapted to control accumulated charge and showing a cross-sectional view line A-A′ taken along approximately a center of the SOI NMOSFET.

FIG. 3F is a cross-sectional view of the improved SOI NMOSET of FIG. 3E taken along the A-A′ view line of FIG. 3E.

FIG. 3F-1 is a cross-sectional view of the improved SOI NMOSET of FIGS. 3A-3B.

FIG. 3G is a simplified top view schematic of an SOI NMOSFET illustrating the region of increased threshold voltage that can occur in the prior art MOSFETs and in some embodiments of the improved SOI MOSFET due to manufacturing processes.

FIG. 3G-1 is a schematic plot of inversion channel charge as a function of applied gate voltage when a region of increased threshold voltage is present.

FIG. 3H is a simplified top view schematic of an improved SOI NMOSFET adapted to control accumulated charge and configured in a “T-gate” configuration.

FIG. 3I is a simplified top view schematic of an improved SOI NMOSFET adapted to control accumulated charge and configured in an “H-gate” configuration.

FIG. 4A is a simplified schematic of an improved SOI NMOSFET adapted to control accumulated charge embodied as a four terminal device.

FIG. 4B is a simplified schematic of an improved SOI NMOSFET adapted to control accumulated charge, embodied as a four terminal device, wherein an accumulated charge sink (ACS) terminal is coupled to a gate terminal.

FIG. 4C is a simplified schematic of an improved SOI NMOSFET adapted to control accumulated charge, embodied as a four terminal device, wherein an accumulated charge sink (ACS) terminal is coupled to a gate terminal via a diode.

FIG. 4D is a simplified schematic of an improved SOI NMOSFET adapted to control accumulated charge, embodied as a four terminal device, wherein an accumulated charge sink (ACS) terminal is coupled to a control circuit.

FIG. 4D-1 is a simplified schematic of the improved SOI NMOSFET of FIG. 4D, embodied as a four terminal device, showing an accumulated charge sink (ACS) terminal coupled to a separate bias source.

FIG. 4E is a simplified schematic of an improved SOI NMOSFET adapted to control accumulated charge, embodied as a four terminal device, wherein an accumulated charge sink (ACS) terminal is coupled to a clamping circuit.

FIG. 4F is a simplified schematic of an improved SOI NMOSFET adapted to control accumulated charge, embodied as a four terminal device, wherein an accumulated charge sink (ACS) terminal is coupled to a gate terminal via a diode in parallel with a capacitor.

FIG. 4G shows plots of the off-state capacitance (C_off) versus applied drain-to-source voltages for SOI MOSFETs operated in the accumulated charge regime, wherein a first plot shows the off-state capacitance C_offof a prior art SOI MOSFET, and wherein a second plot shows the off-state capacitance C_offof the improved ACC SOI MOSFET made in accordance with the present teachings.

FIG. 5A is a schematic of an exemplary single pole, single throw (SPST) radio frequency (RF) switch circuit.

FIG. 5B is a schematic of an RF switch circuit adapted for improved performance using accumulated charge control, wherein the gate of a shunting SOI NMOSFET is coupled to an accumulated charge sink (ACS) terminal.

FIG. 5C is a schematic of an RF switch circuit adapted for improved performance using accumulated charge control, wherein the gate of a shunting SOI NMOSFET is coupled to an accumulated charge sink (ACS) terminal via a diode.

FIG. 5D is a schematic of an RF switch circuit adapted for improved performance using accumulated charge control, wherein the accumulated charge sink (ACS) terminal is coupled to a control circuit.

FIG. 6 is a schematic of an RF switch circuit including stacked MOSFETs, adapted for improved performance using accumulated charge control, wherein the accumulated charge sink (ACS) terminals of the shunting stacked MOSFETs are coupled to a control signal.

FIG. 7 shows a flowchart of an exemplary method of improving the linearity of an SOI MOSFET device using an accumulated charge sink in accordance with the present disclosure.

FIG. 8 shows a simplified circuit schematic of an exemplary embodiment of an RF switch circuit made in accordance with the present disclosure, wherein the RF switch circuit includes drain-to-source resistors between the drain and source of the ACC MOSFETs.

FIG. 9 shows a simplified schematic of an exemplary single-pole double-throw (SPDT) RF switch circuit made in accordance with the present disclosure, wherein drain-to-source resistors are shown across the switching ACC SOI MOSFETs.

FIG. 10A is a flow chart of a first embodiment of a method of implementing SOI MOSFETs to improve gate oxide reliability using the accumulated charge control techniques of the present disclosure.

FIG. 10B is a flow chart of a second embodiment of a method of implementing SOI MOSFETs to improve gate oxide reliability using the accumulated charge control techniques of the present disclosure.

FIG. 10C is a flow chart of a third embodiment of a method of implementing SOI MOSFETs to improve gate oxide reliability using the accumulated charge control techniques of the present disclosure.

Like reference numbers and designations in the various drawings indicate like elements.

DETAILED DESCRIPTION

As noted above, those skilled in the electronic device design arts shall appreciate that the teachings herein apply equally to NMOSFETs and PMOSFETs. For simplicity, the embodiments and examples presented herein for illustrative purposes include only NMOSFETs, unless otherwise noted. By making well known changes to dopants, charge carriers, polarity of bias voltages, etc., persons skilled in the arts of electronic devices will easily understand how these embodiments and examples may be adapted for use with PMOSFETs.

Non-linearity and Harmonic Distortion Effects of Accumulated Charge in an SOI NMOSFET

As described in the background section above, no matter what mode of operation the MOSFET employs (i.e., enhancement mode or depletion mode), under some circumstances, when a MOSFET is operated in an off-state with a nonzero gate bias voltage applied with respect to the source and drain, an accumulated charge may occur under the gate. According to the present teachings, as described above when the MOSFET is in an off-state, and when carriers are present in the channel region having a polarity that is opposite the polarity of the source and drain carriers, the MOSFET is said to be operating in the accumulated charge regime.

According to the present teachings, the inventors have observed that, when used in certain circuit implementations, MOSFETs operating in the accumulated charge regime exhibit undesirable non-linear characteristics that adversely impact circuit performance. For example, as described below in more detail with reference to FIG. 2A, the accumulated charge 120 (FIG. 1) adversely affects the linearity of off-state SOI MOSFETs, and more specifically, it adversely affects the linearity of contributing capacitances to the drain-to-source capacitance (Cds). For an SOI MOSFET operating in an off-state, Cds is referred to as C_off. The contributing capacitances to C _offare described below in reference to FIG. 2A for bias conditions wherein the gate bias Vg is provided by a circuit having an impedance that is large compared to the impedances of the contributing capacitances. As described below with reference to FIGS. 2B and 5A, this, in turn, adversely affects harmonic distortion, intermodulation distortion, and other performance characteristics of circuits implemented with the SOI MOSFETs. These novel observations, not taught or suggested by the prior art, may be understood with reference to the electrical model shown in FIG. 2A.

FIG. 2A is a simplified schematic of an electrical model 200 showing the off-state impedance (or conversely, conductance) characteristics of the exemplary prior art SOI NMOSFET 100 of FIG. 1. More specifically, the model 200 shows the impedance characteristics from the source 112 to the drain 116 when the NMOSFET 100 is operated in the off-state. Because the drain-to-source off-state impedance characteristic of the NMOSFET 100 is primarily capacitive in nature, it is referred to herein as the drain-to-source off-state capacitance (C_off). For the exemplary description herein, the gate 108 is understood to be biased at a voltage Vg by a circuit (not shown) that has an impedance that is large compared to the impedances of the contributing capacitances described in reference to FIG. 2A. Persons skilled in the electronic arts will understand how this exemplary description may be modified for the case wherein the impedance of the circuit providing the Vg bias is not large compared to the impedances of the contributing capacitances.

As shown in FIG. 2A, the junction between the source 112 and the body 114 (i.e., a source-body junction 218) of the off-state NMOSFET 100 can be represented by a junction diode 208 and a junction capacitor 214, configured as shown. Similarly, the junction between the drain 116 and the body 114 (i.e., the drain-body junction 220) of the off-state NMOSFET 100 can be represented by a junction diode 210 and a junction capacitor 216, configured as shown. The body 114 is represented simply as an impedance 212 that is present between the source-body junction 218 and the drain-body junction 220.

A capacitor 206 represents the capacitance between the gate 108 and the body 114. A capacitor 202 represents the capacitance between the source 112 and the gate 108, and another capacitor 204 represents the capacitance between the drain 116 and the gate 108. A substrate capacitance due to the electrical coupling between the source 112 and the drain 116 (through the insulating substrate 118 shown in FIG. 1) is taken to be negligibly small in the exemplary description set forth below, and therefore is not shown in the electrical model 200 of FIG. 2A.

As described above, when the NMOSFET 100 is in the off-state, and when the accumulated charge 120 (FIG. 1) is not present in the body 114 (i.e., the NMOSFET 100 is not operating within the accumulated charge regime), the body 114 is depleted of charge carriers. In this case the body impedance 212 is analogous to the impedance of an insulator, and the electrical conductance through the body 114 is very small (i.e., the NMOSFET 100 is in the off-state). Consequently, the principal contributions to the drain-to-source off-state capacitance C_offare provided by the

capacitors

202 and 204. The

capacitors

202 and 204 are only slightly voltage dependent, and therefore do not significantly contribute to a non-linear response that adversely affects harmonic generation and intermodulation distortion characteristics.

However, when the NMOSFET 100 operates within the accumulated charge regime, and the accumulated charge 120 is therefore present in the body 114, mobile holes comprising the accumulated charge produce p-type conductivity between the source-body junction 218 and the drain-body junction 220. In effect, the accumulated charge 120 produces an impedance between the source-body junction 218 and the drain-body junction 220 that is significantly less than the impedance between the junctions in the absence of the accumulated charge. If a Vds voltage is applied between the drain 116 and the source 112, the mobile holes redistribute according to the electrical potentials that result within the body 114. DC and low-frequency current flow through the SOI NMOSFET 100 is prevented by the diode properties of the source-body junction 218 and the drain-body junction 220, as represented by the

junction diodes

208 and 210, respectively. That is, because the

junction diodes

208 and 210 are anti-series (i.e., “back-to-back”) in this case, no DC or low-frequency currents flow through the SOI NMOSFET 100. However, high-frequency currents may flow through the SOI NMOSFET 100 via the capacitances of the source-body junction 218 and the drain-body junction 220, as represented by the

junction capacitors

214 and 216, respectively.

The

junction capacitors

214 and 216 are voltage dependent because they are associated with junctions between n-type and p-type regions. This voltage dependence results from the voltage dependence of the width of the depletion region of the junction between the n-type and p-type regions. As a bias voltage is applied to the NMOSFET, the width of the depletion region of the junction between the n-type and p-type regions is varied. Because the capacitance of the junction depends on the width of the junction depletion region, the capacitance also varies as a function of the bias applied across the junction (i.e., the capacitance is also voltage dependent).

Further, the

capacitors

202 and 204 may also have a voltage dependence caused by the presence of the accumulated charge 120. Although the complex reasons for this voltage dependence are not described in detail herein, persons skilled in the arts of electronic devices shall understand that electric field regions (e.g.,

electric field regions

122 and 124 described above with reference to FIG. 1) may be affected by the response of the accumulated charge and its response to an applied Vds, thereby causing a voltage dependence of

capacitors

202 and 204. An additional nonlinear effect may occur due to a direct capacitance (not shown) between the source 112 and the drain 116. Although this direct capacitance would usually be expected to be negligible for most SOI MOSFETs, it may contribute for SOI MOSFETs having very short spacing between the source and drain. The contribution of this direct capacitance to C_offis also voltage-dependent in the presence of accumulated charge, for reasons that are analogous to the voltage dependencies of the

capacitors

202 and 204 as described above.

The voltage dependencies of the

junction capacitors

214 and 216, the gate-to-source and gate-to-

drain capacitors

202, 204, respectively, and the direct capacitance (not shown), cause nonlinear behavior in off-state capacitance C_offof the MOSFET when AC voltages are applied to the NMOSFET 100, thereby producing undesirable generation of harmonic distortions and intermodulation distortion (IMD). The relative contributions of these effects are complex, and depend on fabrication processes, biases, signal amplitudes, and other variables. However, those skilled in the electronic device design arts shall understand from the teachings herein that reducing, removing, or otherwise controlling the accumulated charge provides an overall improvement in the nonlinear behavior of C_off. In addition, because the body impedance 212 is significantly decreased in the presence of the accumulated charge 120, the magnitude of C_offmay be increased when the FET operates in the accumulated charge regime. Reducing, removing, or otherwise controlling the accumulated charge also mitigates this effect.

The inventors have observed that the nonlinear behavior of the MOSFET off-state capacitance C_offadversely affects the performance of certain circuits implemented with the prior art SOI MOSFETs. For example, when an RF switch is implemented using the prior art SOI MOSFETs, such as the prior art SOI NMOSFET 100 of FIG. 1, the above-described nonlinear off-state characteristics of the prior art MOSFETs adversely affect the linearity of the switch. As described below in more detail, RF switch linearity is an important design parameter in many applications. Improved switch linearity leads to improved suppression of harmonic and intermodulation (IM) distortion of signals processed by the switch. These improved switch characteristics can be critically important in some applications such as use in cellular communication devices.

For example, the well known GSM cellular communication system standard imposes stringent linearity, harmonic and intermodulation suppression, and power consumption requirements on front-end components used to implement GSM cell phones. One exemplary GSM standard requires that all harmonics of a fundamental signal be suppressed to below −30 dBm at frequencies up to 12.75 GHz. If harmonics are not suppressed below these levels, reliable cell phone operation can be significantly adversely impacted (e.g., increased dropped calls or other communication problems may result due to harmonic and intermodulation distortion of the transmit and receive signals). Because the RF switching function is generally implemented in the cell phone front-end components, improvements in the RF switch linearity, harmonic and intermodulation suppression, and power consumption performance characteristics is highly desirable. A description of how the non-linear behavior of the off-state capacitance C_offof the prior art MOSFETs adversely affects these RF switch characteristics is now described with reference to FIG. 2B.

Harmonic Distortion Effects on RF Switch Circuits Implemented Using Prior Art SOI MOSFETs

FIG. 2B illustrates an exemplary simplified RF switch circuit 250 implemented using prior art MOSFETs such as the prior art SOI NMOSFET 100 described above with reference to FIG. 1. A detailed description of the operation and implementation of RF switch circuits is provided in commonly assigned U.S. Pat. No. 6,804,502 which is hereby incorporated herein by reference in its entirety for its teachings on RF switch circuits. As shown in FIG. 2B, the prior art RF switch 250 includes a single “pass” or “switching” MOSFET 254 operatively coupled to five shunting MOSFETs 260a-260e.

The MOSFET 254 acts as a pass or switching transistor and is configured, when enabled, to selectively couple an RF input signal (applied to its drain, for example) to an RF antenna 258 via a transmission path 256. The shunting MOSFETs, 260a-260e, when enabled, act to alternatively shunt the RF input signal to ground. As is well known, the switching MOSFET 254 is selectively controlled by a first switch control signal (not shown) coupled to its gate, and the shunting MOSFETs, 260a-260e are similarly controlled by a second switch control signal (not shown) coupled to their gates. The switching MOSFET 254 is thereby enabled when the shunting MOSFETs 260a-260e are disabled, and vice versa. As shown in the exemplary embodiment of the RF switch 250 of FIG. 2B, the switching MOSFET 254 is enabled by applying a gate bias voltage of +2.5V (via the first switch control signal). The shunting MOSFETs 260a-260e are disabled by applying a gate bias voltage of −2.5V (via the second switch control signal).

When the switch 250 is configured in this state, the RF signal 252 propagates through the switching MOSFET 254, through the transmission path 256, and to the antenna 258. As described above with reference to FIG. 2A, when the shunting MOSFETS 260a-260e comprise prior art SOI (or SOS) MOSFETs, such as the SOI NMOSFET 100 (FIG. 1), an accumulated charge can occur in the SOI MOSFET bodies (i.e., when the SOI MOSFETs operate in the accumulated charge regime as described above). The accumulated charge can produce nonlinear behavior in the off-state capacitance C_offof the SOI MOSFETs whenAC voltages are applied to the MOSFETs.

More specifically, when the accumulated charge is present in the channel regions of the off-state SOI MOSFETs 260a10 260e it responds to variations in the RF signals applied to their respective drains. As the time varying RF signal propagates along the transmission path 256, the RF signal applies time varying source-to-drain bias voltages to the SOI MOSFETs 260a-260e. The time varying source-to-drain bias voltages creates movement of the accumulated charge within the channel regions of the SOI MOSFETs 260-260e. The movement of the accumulated charge within the channel regions of the SOI MOSFETs causes variations in the drain-to-source off-state capacitance of the SOI MOSFETs 260a-260e. More specifically, the movement of the accumulated charge within the channel regions causes a voltage dependence of the drain-to-source off-state capacitance as described above with reference to FIG. 2A. The voltage dependent variations in the off-state capacitance of the SOI MOSFETs 260a-260e is the dominant cause of harmonic distortion and IMD of the RF signal as it propagates through the RF switch 250.

As noted above, harmonic distortion and IMD of the RF signal is a major disadvantage of the prior art RF switch circuits implemented using the prior art SOI MOSFET devices. For many applications, harmonics and IMD of the RF signal must be suppressed to levels that heretofore have been difficult or impossible to achieve using prior art SOI MOSFET devices. In GSM devices, for example, at a maximum operating power of +35 dBm, prior art switches typically have only a 6 dB margin to the GSM third order harmonics suppression requirement of less than −30 dBm. Very low even order harmonic distortion is also desirable in GSM systems as the second order harmonic of the GSM transmit band also resides in the DCS receive band. Suppression of odd order (e.g., third order) harmonics of the RF signal, however, is desirable and improvements in that regard are needed.

In addition, as is well known, presence of an accumulated charge in the bodies of floating body (e.g., SOI) MOSFETs can also adversely affect the drain-to-source breakdown voltage (BVDSS) performance characteristics of the floating body MOSFETs. As is well known, floating-body FETs demonstrate drain-to-source breakdown voltage problems, also known as BVDSS, wherein the drain-to-source “punch-through” voltage is reduced by a parasitic bipolar action. The parasitic bipolar action is caused when holes are generated in the channel and the holes have nowhere to dissipate (i.e., because the body is floating, the holes have no means for escaping the body). As a consequence, the potential of the MOSFET body is increased, which effectively reduces the threshold voltage. In turn, this condition causes the MOSFET device to experience increased leakage, thereby generating more holes in the body, and thereby exacerbating the BVDSS problem (as a result of this positive feedback condition).

The present disclosed method and apparatus for improving linearity of SOI (and SOS) MOSFET devices overcomes the above-described disadvantages of the prior art. Once the accumulated charge is recognized as a major source of harmonic distortion, IMD and compression/saturation in off-state SOI MOSFET devices, and in circuits (such as RF circuits) implemented with these devices, it becomes clear that reduction, removal, and/or control of the accumulated charge improves the harmonic suppression characteristics of these devices. In addition, reduction, removal, and/or control of the accumulated charge also improves the BVDSS performance characteristics by preventing the parasitic bipolar action from occurring. Improvements in BVDSS lead to consequent improvements in device linearity. Several exemplary structures and techniques for controlling the accumulated charge in SOI MOSFETs are described in detail in the next section.

Method and Apparatus for Improving the Linearity of MOSFETs Using Accumulated Charge Sinks (ACS)—Overview

As described below in more detail, the present disclosure describes methods and apparatuses for improving semiconductor device linearity (e.g., reducing adverse harmonic distortion and IMD effects) in SOI MOSFETs. In one exemplary embodiment, the method and apparatus improves the linearity and controls the harmonic distortion and IMD effects of the MOSFET devices by reducing the accumulated charge in the bodies of the MOSFET devices. In one embodiment, the present method and apparatus reduces or otherwise controls the accumulated charge in the MOSFET bodies using an accumulated charge sink (ACS) that is operatively coupled to the MOSFET body. In one embodiment, the present method and apparatus entirely removes all of the accumulated charge from the bodies of the MOSFET devices. In one described embodiment, the MOSFET is biased to operate in an accumulated charge regime, and the ACS is used to entirely remove, reduce, or otherwise control, the accumulated charge and thereby reduce harmonic distortions and IMD that would otherwise result. Linearity is also improved in some embodiments by removing or otherwise controlling the accumulated charge thereby improving the floating body MOSFET BVDSS characteristics.

As noted in the background section above, persons skilled in the electronic device design and manufacture arts shall appreciate that the teachings herein apply equally to MOSFETs fabricated on Silicon-On-Insulator (“SOI”) and Silicon-On-Sapphire (“SOS”) substrates. As noted above, the present method and apparatus may also be applied to silicon-germanium (SiGe) SOI MOSFETs. For simplicity, the embodiments and examples presented herein for illustrative purposes include only NMOSFETs, unless otherwise noted. By making well known changes to dopants, charge carriers, polarity of bias voltages, etc., persons skilled in the electronic device design arts will easily understand how these embodiments and examples may be adapted for use with PMOSFETs.

As noted above, the present disclosure is particularly applicable to FETs and associated applications benefiting from a fully depleted channel when the FET is operated in the off-state, wherein an accumulated charge may result. The disclosed method and apparatus for use in improving the linearity of MOSFETs also finds applicability for use with partially depleted channels. As known to those skilled in the art, the doping and dimensions of the body vary widely. In an exemplary embodiment, the body comprises silicon having a thickness of approximately 100 angstroms to approximately 2,000 angstroms. In a further exemplary embodiment, dopant concentration within the FET bodies ranges from no more than that associated with intrinsic silicon to approximately 1×10¹⁸active dopant atoms per cm³, resulting in fully-depleted transistor operation. In a further exemplary embodiment, dopant concentration within the FET bodies ranges from 1×10¹⁸to 1×10¹⁹active dopant atoms per cm³and/or the silicon comprising the body ranges from a thickness of 2000 angstroms to many micrometers, resulting in partially-depleted transistor operation. As will be appreciated by those skilled in the electronic design and manufacturing arts, the present disclosed method and apparatus for use in improving linearity of MOSFETs can be used in MOSFETs implemented in wide variety of dopant concentrations and body dimensions. The present disclosed method and apparatus therefore is not limited for use in MOSFETs implemented using the exemplary dopant concentrations and body dimensions as set forth above.

According to one aspect of the present disclosure, accumulated charge within a FET body is reduced using control methodologies and associated circuitry. In one embodiment all of the accumulated charge is removed from the FET body. In other embodiments, the accumulated charge is reduced or otherwise controlled. In one embodiment, holes are removed from the FET body, whereas in another embodiment, electrons are removed from the FET body, as described below in more detail. By removing holes (or electrons) from the FET body using the novel and nonobvious teachings of the present disclosure, voltage induced variations in the parasitic capacitances of the off-state FETs are reduced or eliminated, thereby reducing or eliminating nonlinear behavior of the off-state FETs. In addition, as described above with reference to FIG. 2A, because the body impedance is greatly increased when the accumulated charge is reduced or controlled, there is a beneficial overall reduction in the magnitude of the FET off-state capacitances. Also, as described above, removing or otherwise controlling the accumulated charge in floating body MOSFETs improves the BVDSS characteristics of the FET and thereby improves the linearity of the floating body MOSFET.

Accumulated charge control not only facilitates a beneficial overall reduction in the FET off-state capacitance C_off(as described above with reference to FIG. 2A and below with reference to FIG. 4G), it also facilitates a reduction in C_offvariations that can occur over time in the presence of a time varying V_dsbias voltage. Thus, a reduction of undesirable harmonics generation and intermodulation distortion in RF switch circuits is obtained using SOI MOSFETs made in accordance with the present disclosure. Improved SOI MOSFET power handling, linearity, and performance are achieved by devices made in accordance with the present teachings. While the methods and apparatuses of the present disclosure are capable of fully removing accumulated charge from the FET bodies, those skilled in the electronic device design arts shall appreciate that any reduction of accumulated charge is beneficial.

Reductions in harmonics and intermodulation distortion are generally beneficial in any semiconductor system, either bulk semiconductor or semiconductor-on-insulator (SOI) systems. SOI systems include any semiconductor architecture employing semiconductor-containing regions positioned above an underlying insulating substrate. While any suitable insulating substrate can be used in a SOI system, exemplary insulating substrates include silicon dioxide (e.g., a buried oxide layer supported by a silicon substrate, such as that known as Separation by Implantation of Oxygen (SIMOX)), bonded wafer (thick oxide), glass, and sapphire. As noted above, in addition to the commonly used silicon-based systems, some embodiments of the present disclosure may be implemented using silicon-germanium (SiGe), wherein the SiGe is used equivalently in place of Si.

A wide variety of ACS implementations and structures can be used to practice the present disclosed method and apparatus. In accordance with one embodiment of the present method and apparatus, an ACS is used to remove or otherwise control accumulated charge (referenced as 120 in FIG. 1 described above) from the MOSFETs when the MOSFETs are configured to operate in the accumulated charge regime. By adapting the SOI (or SOS) MOSFETs in accordance with the present teachings, improved Accumulated Charge Control (ACC) MOSFETs are realized. The ACC MOSFETs are useful in improving performance of many digital circuits, including RF switching circuits. Various characteristics and possible configurations of the exemplary ACC MOSFETs are described in detail below with reference to FIGS. 3A-3I. This section also describes how the exemplary ACS implementations of the present disclosure differ from the body contacts of the prior art.

The ACC MOSFET is shown schematically embodied as a four-terminal device in FIG. 4A. FIGS. 4B-4F show various exemplary simple circuit configurations that can be used in removing the accumulated charge from the ACC MOSFET when it operates in an accumulated charge regime. The operation of the simplified circuit configurations is described in more detail below with reference to FIGS. 4A-4F. The improvement in off-state capacitance C_offof the ACC MOSFETs, as compared with the off-state capacitance of the prior art SOI MOSFETs, is described below with reference to FIG. 4G.

The operation of various exemplary RF switch circuits implemented using the ACC MOSFETs of the present disclosure is described below with reference to the circuit schematics of FIGS. 5B-5D. Further, an exemplary RF switch circuit using stacked ACC MOSFETs (for increased power handling) of the present disclosure is described below with reference to FIG. 6. An exemplary method of improving the linearity of an SOI MOSFET using an accumulated charge sink (ACS) is described with reference to FIG. 7. Finally, exemplary fabrication methods that may be used to manufacture the ACC MOSFET are described. The various exemplary ACS implementations and structures that can be used to practice the disclosed method and apparatus are now described with reference to FIGS. 3A-3I.

Controlling Accumulated Charge Using an Accumulated Charge Sink (ACS)

FIGS. 3A and 3B are simplified schematic diagrams of a top view of an Accumulated Charge Control (ACC) SOI NMOSFET 300 adapted to control accumulated charge 120 (FIG. 1) in accordance with the present disclosure. In the exemplary embodiment, a gate contact 301 is coupled to a first end of a gate 302. A gate oxide (not shown in FIG. 3A but shown in FIG. 1) and a body 312 (shown in FIG. 3B) are positioned under the gate 302. In the exemplary NMOSFET 300 shown, a source 304 and a drain 306 comprise N+ regions. In the exemplary embodiment, the ACC NMOSFET 300 includes an accumulated charge sink (ACS) 308 comprising a P− region. The ACS 308 is coupled to and is in electrical communication with the body 312 which also comprises a P− region. A region 310 provides electrical connection to the ACS 308. In some embodiments, the region 310 comprises a P+ region. As shown in FIG. 3A, the region 310 is coupled to and is in electrical communication with the ACS 308.

Those skilled in the arts of electronic devices shall understand that the region 310 may be used to facilitate electrical coupling to the ACS 308 because in some embodiments it may be difficult to make a direct contact to a lightly doped region. In addition, in some embodiments the ACS 308 and the region 310 may be coextensive. In another embodiment, the region 310 comprises an N+ region. In this embodiment, the region 310 functions as a diode connection to the ACS 308, which prevents positive current flow into the ACS 308 (and also prevents positive current flow into the body 312) under particular bias conditions, as described below in more detail.

FIG. 3B is an alternative top view of the ACC SOI NMOSFET 300 of FIG. 3A, illustrating the ACC NMOSFET 300 without its gate contact 301, gate 302, and gate oxide being visible. This view allows the body 312 to be visible. FIG. 3B shows the coupling of the ACS 308 to one end of the body 312. In one embodiment, the body 312 and the ACS 308 comprise a combined P− region that may be produced by a single ion-implantation step. In another embodiment, the body 312 and ACS 308 comprise separate P− regions that are coupled together.

As is well known to those skilled in the electronic device design arts, in other embodiments, the ACC NMOSFET 300 of FIGS. 3A and 3B can be implemented as an ACC PMOSFET simply by reversing the dopant materials used to implement the various FET component regions (i.e., replace p-type dopant material with n-type dopant material, and vice versa). More specifically, in an ACC PMOSFET, the source and drain comprise P+ regions, and the body comprises an N− region. In this embodiment, the ACS 308 also comprises an N− region. In some embodiments of the ACC PMOSFET, the region 310 may comprise an N+ region. In other embodiments of the ACC PMOSFETs, the region 310 comprises a P+ region, which functions as a diode connection to the ACS 308 and thereby prevents current flow into the ACS 308 under particular bias conditions.

Prior Art Body Contacts Distinguished from the Disclosed ACS

According to the present disclosure, the ACS 308 used to implement ACC SOI MOSFETs includes novel features in structure, function, operation and design that distinguish it from the so-called “body contacts” (also sometimes referred to as “body ties”, usually when the “body contact” is directly connected to the source) that are well known in the prior art.

Exemplary references relating to body contacts used in prior art SOI MOSFETs include the following: (1) F. Hameau and O. Rozeau, Radio-Frequency Circuits Integration Using CMOS SOI 0.25 μm Technology,” 2002 RF IC Design Workshop Europe, 19-22 Mar. 2002, Grenoble, France; (2) J. R. Cricci et al., “Silicon on Sapphire MOS Transistor,” U.S. Pat. No. 4,053,916, Oct. 11, 1977; (3) O. Rozeau et al., “SOI Technologies Overview for Low-Power Low-Voltage Radio-Frequency Applications,” Analog Integrated Circuits and Signal Processing, 25, pp. 93-114, Boston, Mass., Kluwer Academic Publishers, November 2000; (4) C. Tinella et al., “A High-Performance CMOS-SOI Antenna Switch for the 2.5-5-GHz Band, “IEEE Journal of Solid-State Circuits, Vol. 38, No. 7, July, 2003; (5) H. Lee et al., “Analysis of body bias effect with PD-SOI for analog and RF applications,” Solid State Electron., Vol. 46, pp. 1169-1176, 2002; (6) J.-H. Lee, et al., “Effect of Body Structure on Analog Performance of SOI NMOSFETs,” Proceedings, 1998 IEEE International SOI Conference, 5-8 Oct. 1998, pp. 61-62; (7) C. F. Edwards, et al., The Effect of Body Contact Series Resistance on SOI CMOS Amplifier Stages,” IEEE Transactions on Electron Devices, Vol. 44, No. 12, December 1997 pp. 2290-2294; (8) S. Maeda, et al., Substrate-bias Effect and Source-drain Breakdown Characteristics in Body-tied Short-channel SOI MOSFET's,” IEEE Transactions on Electron Devices, Vol. 46, No. 1, Jan. 1999 pp. 151-158; (9) F. Assaderaghi, et al., “Dynamic Threshold-voltage MOSFET (DTMOS) for Ultra-low Voltage VLSI,” IEEE Transactions on Electron Devices, Vol. 44, No. 3, March 1997, pp. 414-422; (10) G. O. Workman and J. G. Fossum, “A Comparative Analysis of the Dynamic Behavior of BTG/SOI MOSFETs and Circuits with Distributed Body Resistance,” IEEE Transactions on Electron Devices, Vol. 45, No. 10, October 1998 pp. 2138-2145; and (11) T.-S. Chao, et al., “High-voltage and High-temperature Applications of DTMOS with Reverse Schottky Barrier on Substrate Contacts,” IEEE Electron Device Letters, Vol. 25, No. 2, February 2004, pp. 86-88.

As described herein, applications such as RF switch circuits, may use SOI MOSFETs operated with off-state bias voltages, for which accumulated charge may result. The SOI MOSFETs are defined herein as operating within the accumulated charge regime when the MOSFETs are biased in the off-state, and when carriers having opposite polarity to the channel carriers are present in the channel regions of the MOSFETs. In some embodiments, the SOI MOSFETs may operate within the accumulated charge regime when the MOSFETs are partially depleted yet still biased to operate in the off-state. Significant benefits in improving nonlinear effects on source-drain capacitance can be realized by removing or otherwise controlling the accumulated charge according to the present teachings. In contrast to the disclosed techniques, none of the cited prior art teach or suggest ACS methods and apparatuses that are uniquely useful for removing or controlling accumulated charge. Nor are they informed regarding problems caused by the accumulated charge such as nonlinear effects on the off-state source-drain capacitance C_off. Consequently, the prior art body contacts described in the references cited above differ greatly (in structure, function, operation and design) from the ACSs described with reference to FIGS. 3A-4D.

In one example, the ACS 308 operates effectively to remove or otherwise control the accumulated charge from the SOI NMOSFET 300 using a high impedance connection to and throughout the body 312. High impedance ACSs may be used because the accumulated charge 120 is primarily generated by phenomena (e.g., thermal generation) that take a relatively long period of time to produce significant accumulated charge. For example, a typical time period for producing non-negligible accumulated charge when the NMOSFET operates in the accumulated charge regime is approximately a few milliseconds or greater. Such relatively slow generation of accumulated charge corresponds to very low currents, typically less than 100 nA/mm of transistor width. Such low currents can be effectively conveyed even using very high impedance connections to the body. According to one example, the ACS 308 is implemented with a connection having a resistance of greater than 10⁶ohms. Consequently, the ACS 308 is capable of effectively removing or otherwise controlling the accumulated charge 120 even when implemented with a relatively high impedance connection, relative to the low impedance prior art body contacts.

In stark contrast, the prior art teachings of body contacts described in the references cited above require low impedance (high efficiency) access to the body regions of SOI MOSFETs for proper operation (see, e.g., references (3), (6), and (7) above). A principal reason for this requirement is that the prior art body contacts are primarily directed to reducing the adverse effects on SOI MOSFET functions caused by much faster and more effective electron-hole pair generation processes than occur when the FET is operated in the accumulated charge regime. For example, in some prior art MOSFETs not operated in the accumulated charge regime, electron-hole pair carriers are generated as a result of impact ionization. Impact ionization produces electron-hole pairs at a much faster rate than occurs when the FET is operated in the accumulated charge regime.

The relative rates for electron-hole pair generation by impact ionization versus the pair generation processes causing accumulated charge can be estimated from the roll-off frequencies for the two phenomena. For example, reference (3) cited above indicates roll-off frequencies for impact ionization effects in the range of 10⁵Hz. In contrast, a roll-off frequency for the accumulated charge effects has been observed to be in the range of 10³Hz or less, as indicated by recovery times for odd harmonics. These observations indicate that the ACS 308 can effectively control accumulated charge using an impedance that is at least 100 times larger than required of prior art body contacts used in controlling impact ionization charge, for example. Further, because impact ionization primarily occurs when the SOI MOSFET operates in an on-state, the effects of impact ionization can be amplified by on-state transistor operation. Low impedance body contacts to and throughout a body region is even more critical in these environments in order to control the effects of impact ionization under the on-state conditions.

In stark contrast, the ACS 308 of the present teachings removes or otherwise controls the accumulated charge only when the ACC SOI MOSFET operates in the accumulated charge regime. By definition, the FET is in the off-state in this regime, so there is no requirement to remove impact ionization as amplified by an on-state FET. Therefore, a high impedance ACS 308 is perfectly adequate for removing the accumulated charge under these operating conditions. The prior art requirements for low impedance body connections results in numerous problems of implementation that are overcome by the present teachings, as described below in more detail.

In addition, the ACS 308 may be implemented with much lower source-to-drain parasitic capacitance as compared to the body contacts of the prior art. The above-described low impedance connection to the SOI MOSFET body required of the prior art body contacts necessitates proximity of the contacts to the entire body. This may require a plurality body contact “fingers” that contact the body at different locations along the body. The low impedance connection to the body also necessitates proximity of the prior art body contacts to the source and drain. Because of parasitic capacitances produced by such body contacts, the cited prior art references teach away from the use of such structures for many high frequency applications such as RF. In stark contrast, the ACS 308 of the present disclosure may be positioned away from the source 304, the body 312, and the drain 306, as shown in FIGS. 3A and 3B, thereby eliminating or greatly reducing parasitic capacitances caused by a more proximate positioning of the ACS 308 relative to the source drain and body. Further, the ACS 308 may be implemented in SOI MOSFETs operated with a depleted channel. In general, the cited prior art references teach away from the use of body contacts for this environment (see, e.g., reference (3), cited above).

Further, because impact ionization hole currents are much larger (in the range of 5,000 nA per mm body width) than for accumulated charge generation (less than approximately 100 nA per mm body width), the prior art does not teach how to effectively implement very large body widths (i.e., much greater than approximately 10 μm). In contrast, the ACS 308 of the present disclosed device may be implemented in SOI MOSFETs having relatively large body widths. This provides improvements in on-state conductance and transconductance, insertion loss and fabrication costs, particularly for RF switch devices. According to the prior art teachings cited above, larger body widths adversely affect the efficient operation of body contacts because their impedances are necessarily thereby increased. Although the cited prior art suggests that a plurality of fingers may be used to contact the body at different locations, the plurality of fingers adversely affects parasitic source-to-drain capacitances, as described above.

For these reasons, and for the reasons described below in more detail, the present disclosure provides novel MOSFET devices, circuits and methods that overcome the limitations according to the prior art teachings as cited above.

FIG. 3C is a cross-sectional perspective schematic of an ACC SOI NMOSFET 300′ adapted to control accumulated charge in accordance with the disclosed method and apparatus. In the example shown in FIG. 3C, the ACC NMOSFET 300′includes four terminals that provide electrical connection to the various FET component regions. In one embodiment, the terminals provide means for connecting external integrated circuit (IC) elements (such as metal leads, not shown) to the various FET component regions. Three of the terminals shown in FIG. 3C are typically available in prior art FET devices. For example, as shown in FIG. 3C, the ACC NMOSFET 300′ includes a gate terminal 302′ that provides electrical connection to the gate 302. Similarly, the ACC NMOSFET 300′ includes source and drain terminals 304′, 306′ that provide electrical connection to the source 304 and drain 306, respectively. As is well known in the electronic design arts, the terminals are coupled to their respective FET component regions (i.e., gate, drain and source) via so-called “ohmic” (i.e., low resistance) contact regions. The manufacturing and structural details associated with the coupling of the various FET terminals to the FET component regions are well known in the art, and therefore are not described in more detail here.

As described above with reference to FIGS. 3A and 3B, the ACC NMOSFET 300′ is adapted to control accumulated charge when the NMOSFET operates in the accumulated charge regime. To this end, in the exemplary embodiment shown in FIG. 3C, the ACC NMOSFET 300′ includes a fourth terminal that provides electrical connection to the body 312, and thereby facilitates reduction (or other control) of the accumulated charge when the FET 300′ operates in the accumulated charge regime. More specifically, and referring again to FIG. 3C, the ACC NMOSFET includes a “body” terminal, or Accumulated Charge Sink (ACS) terminal 308′. The ACS terminal 308′ provides an electrical connection to the ACS 308 (not shown in FIG. 3C, but shown in FIGS. 3A and 3B) and to the body 312. Although the ACS terminal 308′ is shown in FIG. 3C as being physically coupled to the body 312, those skilled in the electronic design arts shall understand that this depiction is for illustrative purposes only. The direct coupling of the ACS terminal 308′ to the body 312 shown in FIG. 3C illustrates the electrical connectivity (i.e., not the physical coupling) of the terminal 308′ with the body 312. Similarly, the other terminals (i.e., terminals 302′, 304′ and 306′) are also shown in FIG. 3C as being physically coupled to their respective FET component regions. These depictions are also for illustrative purposes only.

In most embodiments, as described above with reference to FIGS. 3A-3B, and described further below with reference to FIGS. 3D-3I, the ACS terminal 308′ provides the electrical connection to the body 312 via coupling to the ACS 308 via the region 310. However, the present disclosure also contemplates embodiments where the coupling of the ACS terminal 308′ is made directly to the body 312 (i.e., no intermediate regions exist between the ACS terminal and the body).

In accordance with the disclosed method and apparatus, when the ACC NMOSFET 300′ is biased to operate in the accumulated charge regime (i.e., when the ACC NMOSFET 300′ is in the off-state, and there is an accumulated charge 120 of P polarity (i.e., holes) present in the channel region of the body 312), the accumulated charge is removed or otherwise controlled via the ACS terminal 308′. When accumulated charge 120 is present in the body 312, the charge 312 can be removed or otherwise controlled by applying a bias voltage (V_b(for “body”) or V_ACS(ACS bias voltage)) to the ACS terminal 308′. In general, the ACS bias voltage V_ACSapplied to the ACS terminal 308′ may be selected to be equal to or more negative than the lesser of the source bias voltage Vs and drain bias voltage Vd. More specifically, in some embodiments, the ACS terminal 308′ can be coupled to various accumulated charge sinking mechanisms that remove (or “sink”) the accumulated charge when the FET operates in the accumulated charge regime. Several exemplary accumulated charge sinking mechanisms and circuit configurations are described below with reference to FIGS. 4A-5D.

Similar to the prior art NMOSFET 100 described above with reference to FIG. 1, the ACC SOI NMOSFET 300′ of FIG. 3C can be biased to operate in the accumulated charge regime by applying specific bias voltages to the various terminals 302′, 304′, and 306′. In one exemplary embodiment, the source and drain bias voltages (Vs and Vd, respectively) are zero (i.e., the terminals 304′ and 306′ are connected to ground). In this example, if the gate bias voltage (Vg) applied to the gate terminal 302′ is sufficiently negative with respect to the source and drain bias voltages, and with respect to V_th(for example, if V_this approximately zero, and if Vg is more negative than approximately −1 V), the ACC NMOSFET 300′ operates in the off-state. If the ACC NMOSFET 300′ continues to be biased in the off-state, the accumulated charge (holes) will accumulate in the body 312. Advantageously, the accumulated charge can be removed from the body 312 via the ACS terminal 308′. In some embodiments, as described below in more detail with reference to FIG. 4B, the ACS terminal 308′ is coupled to the gate terminal 302′ (thereby ensuring that the same bias voltages are applied to both the gate (Vg) and the body (shown in FIG. 3C as “Vb” or “V_ACS”).

However, those skilled in the electronics design arts shall appreciate that a myriad of bias voltages can be applied to the four device terminals while still employing the techniques of the present disclosed method and apparatus. As long as the ACC SOI NMOSFET 300′ is biased to operate in the accumulated charge regime, the accumulated charge can be removed or otherwise controlled by applying a bias voltage V_ACSto the ACS terminal 308′, and thereby remove the accumulated charge from the body 312.

For example, in one embodiment wherein the ACC NMOSFET 300′ comprises a depletion mode device, V_this negative by definition. In this embodiment if both the Vs and Vd bias voltages comprise zero volts (i.e., both terminals tied to circuit ground node), and a gate bias Vg applied to the gate terminal 302′ is sufficiently negative to V_th(for example, Vg is more negative than approximately −1 V relative to V_th), holes may accumulate under the gate oxide 110 thereby becoming the accumulated charge 120. In this example, in order to remove the accumulated holes (i.e., the accumulated charge 120) from the FET body 312, the voltage V_ACSapplied to the ACS 308 may be selected to be equal to or more negative than the lesser of Vs and Vd.

In other examples, the source and drain bias voltages, Vs and Vd, respectively, may comprise voltage other than zero volts. According to these embodiments, the gate bias voltage Vg must be sufficiently negative to both Vs and Vd (in order for Vg to be sufficiently negative to V_th, for example) in order to bias the NMOSFET in the off-state. As described above, if the NMOSFET is biased in the off-state for a sufficiently long time period (approximately 1-2 ms, for example) an accumulated charge will accumulate under the gate oxide. In these embodiments, as noted above, in order to remove the accumulated charge 120 from the body 312, the ACS bias voltage V_ACSapplied to the ACS terminal 308′ may be selected to be equal to or more negative than the lesser of Vs and Vd.

It should be noted that, in contrast to the examples described above, the prior art body contacts are implemented largely for purposes of mitigating the adverse effects caused by impact ionization. Consequently, the prior art body contacts are typically tied to the source of the MOSFET. In order to effectively control, reduce, or entirely remove the accumulated charge in an NMOSFET, V_ACSshould, in the exemplary embodiments, be equal to or more negative than the lesser of Vs and Vd. Those skilled in the electronic device design arts shall appreciate that different Vs, Vd, Vg and V_ACSbias voltages may be used when the ACC MOSFET comprises a PMOSFET device. Because the prior art body contacts are typically tied to the source, this implementation cannot be effected using the prior art body contact approach.

FIG. 3D is a simplified schematic diagram of a top view of an ACC SOI NMOSFET 300″ adapted to control accumulated charge 120 (FIG. 1) in accordance with the present disclosure. FIG. 3D shows the ACC NMOSFET 300″ without its gate contact 301, gate 302, and gate oxide being visible. The ACC NMOSFET 300″ of FIG. 3D is very similar in design to the ACC NMOSFET 300 described above with reference to FIGS. 3A and 3B. For example, similar to the ACC NMOSFET 300, the ACC NMOSFET 300″ includes a source 304 and drain 306 comprising N+ regions. The ACC NMOSFET 300″ also includes an accumulated charge sink (ACS) 308 comprising a P− region. As shown in FIG. 3D, the P− region that comprises the ACS 308 abuts (i.e., is directly adjacent) the body 312, which also comprises a P− region. Similar to the ACC NMOSFET 300, the ACC NMSOFET 300″ includes a region 310 that provides electrical connection to the ACS 308. As noted above, in some embodiments, the region 310 comprises a P+ region. In another embodiment, the region 310 may comprise an N+ region (which thereby prevents positive current flow into the body 312 as noted above). As shown in FIG. 3D, the region 310 is formed in the ACC NMOSFET 300″ directly adjacent the ACS 308. The ACC SOI NMOSFET 300″ functions to control accumulated charge similarly to the operation of the ACC NMOSFETs described above with reference to FIGS. 3A-3C.

FIG. 3E is a simplified schematic diagram of a top view of an ACC SOI NMOSFET 300′ adapted to control accumulated charge in accordance with the present disclosure. The ACC NMOSFET 300′″ is very similar in design and function to the ACC NMOSFETs described above with reference to FIGS. 3A-3D. FIG. 3E shows a dashed cross-sectional view line A-A′ taken along the approximate center of the NMOSFET 300′″. This cross-sectional view is used herein to describe structural and performance characteristics of some exemplary prior art MOSFETS and some embodiments of the ACC NMOSFET that may occur as a result of the fabrication processes. Details of this cross-sectional view A-A′ are now described with reference to FIG. 3F.

View line A-A′ slices through the following component regions of the ACC NMOSFET 300′″: the P+ region 310, the ACS 308 (shown in FIG. 3E, but not shown in FIG. 3F), a P+ overlap region 310′, a gate oxide 110, and a poly-silicon gate 302. In some embodiments, during the fabrication process, when the region 310 is doped with p-type dopant material, proximate the P− body region, some additional P+ doping may be implanted (i.e., the p-type dopant material may overlap) into the P+ overlap region 310′ of the poly-silicon gate 302. In some embodiments, such overlapping is intentionally performed to ensure that all of the gate oxide 110 is completely covered by the P+ region (i.e., to ensure that no gap exists on the edge of the oxide 110 between the gate 302 and the P+ region 310). This, in turn, aids in providing a minimum impedance connection between the P+ region 310 and the body 312.

Although the present teachings encompass such embodiments described above, those skilled in the electronic device design and manufacturing arts shall recognize that such low-resistance connections are not required. Therefore, the disadvantages associated with the embodiment shown in FIGS. 3G and 3G-1, as described below in more detail, can be overcome by using other embodiments described herein (for example, the

embodiments

300 and 300″″ described below with reference to FIGS. 3F-1 and 3H, respectively), in which gaps are intentionally implemented between the P+ region 310 and the body 312. In one exemplary embodiment, the P+ overlap region 320 overlaps the oxide 110 by approximately 0.2-0.7 microns. Those skilled in the MOSFET design and manufacturing arts shall appreciate that other overlap region dimensions can be used in practicing the present disclosed method and apparatus. In some embodiments, as shown in FIG. 3F, the remaining area over the gate oxide 110 and over the P− body is doped with n-type dopant material (i.e., it comprises an N+ region).

Owing to the presence of the P+ overlap region 310′ over the gate oxide 110, over the body 312, and proximate an edge 340 of the poly-silicon gate 302, an increased threshold voltage is created in one region of the NMOSFET 300′″. More specifically, due to the P+ doping (in the overlap region 310′) proximate the edge 340 of the gate 302 over the channel region of the body 312, a higher magnitude threshold voltage is created in that portion of the MOSFET 300′″. The effects of the increased threshold voltage are now described in more detail with reference to FIG. 3G.

FIG. 3G-1 is a schematic plot of the inversion channel charge vs. applied gate voltage of an ACC NMOSFET illustrating one effect of the above described increased threshold voltage that can occur in prior art MOSFETs and in some embodiments of the present ACC NMOSFETS due to manufacturing processes. As described below in more detail, the increased magnitude threshold voltage shown in FIG. 3G also occurs in prior art designs because of the proximity of body ties to the FET body. As described below with reference to the embodiment of FIG. 3H, for example, the present disclosed method and apparatus can be used to reduce or eliminate the increased threshold voltage found in the prior art SOI MOSFET designs.

FIG. 3G shows one embodiment of the ACC NMOSFET without its gate contact, gate, and gate oxide being visible. The MOSFET region of increased threshold voltage described above with reference to FIGS. 3E and 3F is shown in FIG. 3G as occurring in the region encompassed by the ellipse 307. As will be well understood by those skilled in the electronic design and manufacturing arts, for the reasons set forth above with reference to FIGS. 3E and 3F, due to the increased threshold voltage, the region 307 of the ACC MOSFET shown in FIG. 3G effectively inverts after the rest of the ACC MOSFET channel region.

The threshold voltage increase can be reduced by reducing the size of the region 307. Eliminating the region 307 will eliminate the threshold voltage increase. Because the threshold voltage increase can increase harmonic and intermodulation distortion of the “on” MOSFET, eliminating this effect improves device performance.

In one exemplary embodiment, as shown, for example in the embodiments of the ACC NMOSFET 300 described above with reference to FIGS. 3A and 3B, and as described below in more detail with reference to the cross-sectional view of the ACC MOSFET 300 of FIG. 3F-1, the detrimental effects associated with threshold voltage increase are mitigated or overcome by positioning the P+ region 310 a selected distance away from an edge of the poly-silicon gate 302. This approach is shown both in the top view of the ACC MOSFET 300 of FIG. 3A, and in the cross-sectional view of the ACC MOSFET 300 shown in FIG. 3F-1. As shown in the cross-sectional view of the ACC MOSFET 300 of FIG. 3F-1, the P+ region 310 does not extend all the way to the edge 340 of the poly-silicon gate 302. This is in stark contrast to the embodiment 300′″ shown in FIG. 3F, where the P+ region 310′ extends all the way to the gate edge 340. By positioning the P+ region 310 a distance away from the gate edge 340 as shown in the embodiment 300 of FIG. 3F-1, no P+ region is positioned proximate the poly-silicon gate 302 (i.e., there is no P+ region present in the poly-silicon gate 302). This configuration of the P+ region 310 eliminates or greatly reduces the threshold voltage increase problems described above. As described above with reference to FIGS. 3A and 3B, and with reference to the comparisons to the prior art body contact references, the relatively high impedance of the ACS 308 P− region (shown in FIG. 3A) between the P+ region 310 and the gate 302 does not adversely affect the performance of the ACC NMOSFET 300. As described above, the accumulated charge can be effectively removed even using a relatively high impedance ACS connection.

In another exemplary embodiment, as described below with reference to FIG. 3H, threshold voltage increase is removed by positioning the P+ region 310 (and the ACS 308) a distance away from the body 312. Because the electrical connectivity between the ACS 308 and the body 312 has relatively high impedance when the small region of P+ 310 is positioned a distance away from the body 312, this approach is never taught or suggested by the body contact prior art references (which require low impedance contacts as described above). This improved embodiment is described next with reference to FIG. 3H.

FIG. 3H is a simplified top view schematic of another embodiment of an ACC SOI NMOSFET 300″″ adapted to control accumulated charge and configured in a “T-gate” configuration. FIG. 3H shows the ACC NMOSFET 300″″ without its gate contact 301, gate 302, and gate oxide being visible. As shown in FIG. 3H, the ACC NMOSFET 300″″ includes a small P+ region 310 conjoined to an ACS 308. As shown in FIG. 3H, the P+ region 310 (and thus the ACS external electrical connection) is disposed a selected distance away from the body 312. The total impedance of the electrical connection from the body 312, through the ACS 308, and to the P+ region 310 is increased by positioning the P+ region 310 a selected distance away from the body 312. However, as described above, the present ACC NMOSFET 300″″ works perfectly well to remove accumulated charge even using relatively high impedance ACS connections. For the reasons described above with reference to FIGS. 3A and 3B, due to the nature of the accumulated charge when the NMOSFET 300″″ operates in the accumulated charge regime, the ACC NMOSFET 300″″ does not require low impedance ACS electrical connections in order to remove accumulated charge from the body 312. Rather, an ACS connection of relatively large impedance may be used in practicing the present teachings, with corresponding improvements in NMOSFET performance as described above (e.g., reductions in parasitic capacitance as compared with prior art low impedance body contacts). However, in other embodiments, if desired, a low impedance ACS connection may be used to practice the disclosed method and apparatus for use in improving linearity characteristics of SOI MOSFETs.

Moreover, as described above with reference to FIG. 3G, the embodiment of FIG. 3H improves device performance owing to the fact that the small P+ region 310 is positioned a distance away from the body 312. Because the small P+ region 310 is positioned a distance away from the body 312, the threshold voltage increase is reduced or entirely eliminated, together with the consequent adverse performance effects described above.

FIG. 3I is a simplified top view schematic of another embodiment of an ACC SOI NMOSFET 300′″″ adapted to control accumulated charge and configured in an “H-gate” configuration. FIG. 3I shows the ACC NMOSFET 300′″″ without its gate contact 301, gate 302, and gate oxide being visible. With the exception of some structural differences described herein, the ACC NMOSFET 300′″″ is very similar in design and function to the ACC NMOSFETs described above with reference to FIGS. 3A-3D and 3H. As shown in FIG. 3I, the ACC NMOSFET 300′″″ includes two ACSs, 308 and 308″, disposed at opposite ends of the H-gate ACC NMOSFET 300′″″.

P+ regions

310 and 310′ are formed to abut their respective ACSs, 308 and 308″, and provide electrical contact thereto. In accordance with the disclosed method and apparatus, as described above, when the ACC NMOSFET 300′″″ is biased to operate in the accumulated charge regime, the accumulated charge is removed or otherwise controlled via the two ACSs 308 and 308″.

It shall be understood by those skilled in the electronic device design arts that although the illustrated embodiment shows the

ACSs

308 and 308″ extending approximately the entire width of the ACC NMOSFET 300′″″, the

ACSs

308 and 308″ may also comprise much narrower (or wider) regions, and still function perfectly well to remove or otherwise control the accumulated charge. Also, in some embodiments, it is not necessary that the impedance of the ACS 308 matches the impedance of the ACS 308″. It will further be understood by the skilled person that the

ACSs

308 and 308″ may comprise different sizes and configurations (i.e., rectangular, square, or any other convenient shape), and may also be positioned at various distances away from the body 312 (i.e., not necessarily the same distance away from the body 312). As described above with reference to FIG. 3H, when the ACS 308 is positioned a selected distance away from the body 312, the problems associated with threshold voltage increase are reduced or eliminated.

Four-Terminal ACC MOSFET Devices—Simple Circuit Configurations

The SOI NMOSFET 300 of FIGS. 3A and 3B may be implemented as a four terminal device, as illustrated schematically in FIG. 4A. As shown in the improved ACC SOI NMOSFET 300 of FIG. 4A, a gate terminal 402 is electrically coupled to the gate contact 301 (e.g., FIG. 3A) and is analogous to the gate terminal 302′ shown in FIG. 3C. The gate contact 301 is electrically coupled to the gate 302 (e.g., FIGS. 3A and 3C). Similarly, a source terminal 404 is electrically coupled to the source 304 (e.g., FIGS. 3A-3C) and is analogous to the source terminal 304′ of FIG. 3C. Similarly, a drain terminal 406 is electrically coupled to the drain 306 (e.g., FIGS. 3A-3C) and is analogous to the drain terminal 306′ of FIG. 3C. Finally, the ACC NMOSFET 300 includes an ACS terminal 408 that is electrically coupled to the ACS 308 (e.g., see FIGS. 3A-3B, and FIGS. 3D, 3H-3I) via the region 310. Those skilled in the electronic design and manufacturing arts shall understand that the region 310 may be used in some embodiments to facilitate electrical coupling to the ACS 308 because, in some embodiments, it may be difficult to make a direct contact to a lightly doped region (i.e., the ACS 308). The ACS terminal 408 is analogous to the ACS terminal 308′ shown in FIG. 3C.

The ACC SOI NMOSFET 300 of FIG. 4A may be operated using various techniques and implemented in various circuits in order to control accumulated charge present in the FET when it is operating in an accumulated charge regime. For example, in one exemplary embodiment as shown in FIG. 4B, the gate and ACS terminals, 402 and 408, respectively, are electrically coupled together. In one embodiment of the simplified circuit shown in FIG. 4B, the source and drain bias voltages applied to the

terminals

404 and 406, respectively, may be zero. If the gate bias voltage (Vg) applied to the gate terminal 402 is sufficiently negative with respect to the source and drain bias voltages applied to the

terminals

404 and 406, and with respect to the threshold voltage V_th, (for example, if V_this approximately zero, and if Vg is more negative than approximately −1 V) the ACC NMOSFET 300 operates in the accumulated charge regime. As described above with reference to FIG. 3C, for example, when the MOSFET operates in this regime, accumulated charge (holes) may accumulate in the body of the NMOSFET 300.

Advantageously, the accumulated charge can be removed via the ACS terminal 408 by connecting the ACS terminal 408 to the gate terminal 402 as shown. This configuration ensures that when the FET 300 is in the off-state, it is held in the correct bias region to effectively remove or otherwise control the accumulated charge. As shown in FIG. 4B, connecting the ACS terminal 408 to the gate ensures that the same bias voltages are applied to both the gate (Vg) and the body (shown in FIG. 3C as “Vb” or “V_ACS”). Because the bias voltage V_ACSis the same as the gate voltage Vg in this embodiment, the accumulated charge is no longer trapped below the gate oxide (by attraction to the gate bias Vg) because it is conveyed to the gate terminal 402 via the ACS terminal 408. The accumulated charge is thereby removed from the body via the ACS terminal 408.

In other exemplary embodiments, as described above with reference to FIG. 3C, for example, Vs and Vd may comprise nonzero bias voltages. According to these examples, Vg must be sufficiently negative to both Vs and Vd in order for Vg to be sufficiently negative to V_thto turn the NMOSFET 300 off (i.e., operate the NMOSFET 300 in the off-state). When so biased, as described above, the NMOSFET 300 may enter the accumulated charge regime and thereby have accumulated charge present in the body. For this example, the voltage V_ACSmay also be selected to be equal to Vg by connecting the ACS terminal 408 to the gate terminal 402, thereby conveying the accumulated charge from the body of the ACC NMOSFET, as described above.

In another exemplary embodiment, as described above, the ACC NMOSFET 300 comprises a depletion mode device. In this embodiment, the threshold voltage, V_this, by definition, less than zero. For Vs and Vd both at zero volts, when a gate bias Vg sufficiently negative to V_this applied to the gate terminal 402 (for example, Vg more negative than approximately −1 V relative to V_th), holes may accumulate under the gate oxide and thereby comprise an accumulated charge. For this example, the voltage V_ACSmay also be selected to be equal to Vg by connecting the ACS terminal 408 to the gate terminal 402, thereby conveying the accumulated charge from the ACC NMOSFET as described above.

In some embodiments of the improved ACC SOI NMSOFET 300, such as that described above with reference to FIG. 4B, when the FET is biased on, diodes formed at the edge of the device (such as described above with reference to the interface between the ACS 308 and the drain 304 (and the source 306) as shown in FIG. 3D) may become forward biased thereby allowing current to flow into the source and drain regions. In addition to wasting power, this may introduce nonlinearity into the NMSOFET. The nonlinearity results because the current that flows as a result of the forward biased interface diodes comprises nonlinear current. As Vgs and Vgd are reduced in that region of the device, the on resistance Ron at the edge of the device is increased. As is well known, and for the reasons set forth above, if the interface diodes formed at the edge of the device become forward biased, the device on-state characteristics are consequently dramatically adversely affected. Those skilled in the electronic device design arts shall understand that the configuration shown in FIG. 4B limits application of a gate bias voltage Vgs to approximately 0.7 Volts. The simplified circuit shown in FIG. 4C can be used to overcome these problems.

Another exemplary simplified circuit using the improved ACC SOI NMSOFET 300 is shown in FIG. 4C. As shown in FIG. 4C, in this embodiment, the ACS terminal 408 may be electrically coupled to a diode 410, and the diode 410 may, in turn, be coupled to the gate terminal 402. This embodiment may be used to prevent a positive current flow into the MOSFET body 312 caused by a positive Vg-to-Vs (or, equivalently, Vgs, where Vgs=Vg−Vs) bias voltage, as may occur, for example, when the SOI NMOSFET 300 is biased into an on-state condition.

As with the device shown in FIG. 4B, when biased off, the ACS terminal voltage V_ACScomprises the gate voltage plus a voltage drop across the diode 410. At very low ACS terminal current levels, the voltage drop across the diode 410 typically also is very low (e.g., <<100 mV, for example, for a typical threshold diode). The voltage drop across the diode 410 can be reduced to approximately zero by using other diodes, such as a 0Vf diode, for example. In one embodiment, reducing the voltage drop across the diode is achieved by increasing the diode 410 width. Additionally, maintaining the ACS-to-source or ACS-to-drain voltage (whichever bias voltage of the two bias voltages is lower) increasing negative, also improves the linearity of the ACC MOSFET device 300.

When the SOI NMOSFET 300 is biased in an on condition, the diode 410 is reverse-biased, thereby preventing the flow of positive current into the source and drain regions. The reverse-biased configuration reduces power consumption and improves linearity of the device. The circuit shown in FIG. 4C therefore works well to remove accumulated charge from the ACC MOSFET body when the FET is in the off-state and is operated in the accumulated charge regime. It also permits almost any positive voltage to be applied to the gate voltage Vg. This, in turn, allows the ACC MOSFET to effectively remove accumulated charge when the device operates in the off-state, yet assume the characteristics of a floating body device when the device operates in the on-state.

With the exception of the diode 410 used to prevent the flow of positive current into the ACS terminal 408, exemplary operation of the simplified circuit shown in FIG. 4C is the same as the operation of the circuit described above with reference to FIG. 4B.

In yet another embodiment, the ACS terminal 408 may be coupled to a control circuit 412 as illustrated in the simplified circuit of FIG. 4D. The control circuit 412 may provide a selectable ACS bias voltage V_ACSthat selectively controls the accumulated charge (i.e., the accumulated charge 120 described above with reference to FIG. 1). As shown in FIG. 4D, rather than having a local circuit provide the ACS bias voltage V_ACS(e.g., as derived from the gate voltage Vg), in some embodiments the ACS bias voltage V_ACSis produced by a separate source that is independent of the ACC MOSFET device 300. In the case of a switch (as described below in more detail with reference to FIG. 4D-1), the ACS bias voltage V_ACSshould be driven from a source having a high output impedance. For example, such a high output impedance source can be obtained using a large series resistor in order to ensure that the RF voltage is divided across the MOSFET and that the ACS bias voltage V_ACShas Vds/2 “riding” on it, similarly to the gate voltage. This approach is described in more detail below with reference to FIG. 4D-1.

It may be desirable to provide a negative ACS bias voltage V_ACSto the ACS terminal 408 when the SOI NMOSFET 300 is biased into an accumulated charge regime. In this exemplary embodiment, the control circuit 412 may prevent positive current flow into the ACS terminal 408 by selectively maintaining an ACS bias voltage V_ACSthat is consistently negative with respect to both the source and drain bias voltages. In particular, the control circuit 412 may be used to apply an ACS bias voltage that is equal to or more negative than the lesser of Vs and Vd. By application of such an ACS bias voltage, the accumulated charge is thereby removed or otherwise controlled.

In the exemplary embodiment of the simplified circuit shown in FIG. 4D, the source and drain bias voltages applied to the

terminals

404 and 406, and with respect to V_th, (for example, if V_this approximately zero, and if Vg is more negative than approximately −1 V) the ACC NMOSFET 300 operates in the accumulated charge regime, and the accumulated charge (holes) may accumulate in the body of the ACC NMOSFET 300. Advantageously, the accumulated charge can be removed via the ACS terminal 408 by connecting the ACS terminal 408 to the control circuit 412 as shown. In order to ensure that the accumulated charge is conveyed from the body of the ACC NMOSFET 300, the ACS bias voltage V_ACSthat is applied to the ACS terminal 408 should be equal to or more negative than the gate voltage the lesser of Vs and Vd. Because the accumulated charge 120 is conveyed to the bias voltage V_ACSapplied to the ACS terminal 408 by the control circuit 412, the accumulated charge does not remain trapped under the gate oxide due to attraction to the gate bias voltage Vg.

In other embodiments, Vs and Vd may comprise bias voltages that are other than zero. According to these examples, Vg must be sufficiently negative to both Vs and Vd in order for Vg to be sufficiently negative to V_th, in order to bias the NMOSFET 300 in the off-state. This allows the accumulation of accumulated charge under the gate oxide. For this example, the ACS bias voltage V_ACSmay be selected to be equal to or more negative than the lesser of Vs and Vd by connecting the ACS terminal 408 to the control circuit 412 to provide selected ACS bias voltages, thereby conveying the accumulated charge from the ACC NMOSFET 300.

In other embodiments, if the ACC NMOSFET 300 of FIG. 4D comprises a depletion mode device, V_this, by definition, less than zero. For Vs and Vd both at zero volts, when a gate bias Vg sufficiently negative to V_this applied (for example, Vg more negative than approximately −1 V relative to V_th), holes may accumulate under the gate oxide. For this example, the ACS bias voltage V_ACSthat is applied to the ACS terminal 408 may also be selected to be equal to or more negative than the lesser of Vs and Vd by connecting the ACS terminal 408 to the control circuit 412 and thereby provide the desired ACS bias voltages V_ACSthat are necessary to remove the accumulated charge from the ACC NMOSFET 300.

As described above, in one embodiment, instead of having the control circuit 412 provide a bias to the ACS terminal 408 as shown in FIG. 4D, the ACS terminal 408 can be driven by a separate bias source circuit, as shown, for example, in the embodiment of FIG. 4D-1. In one exemplary circuit implementation, as exemplified in the circuit of FIG. 4D-1, in an RF switch circuit, the separate V_ACSsource has a high output impedance element 403 which ensures that the RF voltage is divided across the ACC NMOSFET 300, and which further ensures that the voltage applied to the ACS terminal 408 has Vds/2 applied thereon, similar to the voltage Vgs that is applied to the gate terminal 402. In one exemplary embodiment, an inverter 405 is configured in series with the high output impedance element 403 and supplied by GND and −V_DD. In one exemplary embodiment, −V_DDis readily derived from a convenient positive voltage supply. It could, however, comprise an even more negative voltage for improved linearity (i.e., it can be independent of the gate voltage).

In another embodiment, the circuit shown in FIG. 4C can be modified to include a clamping circuit configured in series with an ACS terminal 408. Such an exemplary embodiment is shown in FIG. 4E. Under certain operating conditions, current that flows out of the ACC NMOSFET 300, conveying the accumulated charge from the body of the ACC NMOSFET 300, via the ACS terminal 408 is sufficiently high such that it causes problems in the biasing circuitry (i.e., under some conditions the ACS current is so high that the biasing circuitry cannot adequately sink the current flowing out of the body of the ACC NMOSFET 300). As shown in the circuit of FIG. 4E, one exemplary embodiment solves this problem by interrupting the flow of ACS current out of the body of the ACC NMOSFET 300, and thereby returning the ACC NMOSFET 300 to a floating body condition.

In one exemplary circuit, as shown in FIG. 4E, a depletion-mode FET 421 is configured in series between the ACS terminal 408 and a diode 410. In this exemplary circuit, the depletion-mode FET 421 includes a gate terminal that is electrically connected to the FET's source terminal. In this configuration, the depletion-mode FET 421 functions to clip or limit the current that flows from the ACS terminal 408 when the ACC MOSFET operates in the accumulated charge regime. More specifically, the depletion-mode FET 421 enters saturation upon reaching a predefined threshold. The current leaving the body of the ACC MOSFET is thereby limited by the saturation current of the FET 421. In some embodiments, the predefined saturation threshold may optionally be adjusted to change the point at which clamping occurs, such as by selecting a higher threshold voltage, which results in a lower maximum current and earlier clamping.

In some embodiments, such as for example in an RF switch circuit, the gate terminal 402 and the ACS terminal 408 follow Vds at half the value (Vds/2) of Vds. At high Vds excursions, Vgs may approach the threshold voltage Vth, resulting in increased Ids leakage current. In some cases, such a leakage current exits the ACS terminal 408 and can overwhelm associated circuitry (e.g., a negative voltage generator). Hence, the circuit shown in FIG. 4E solves or otherwise mitigates these problems. More specifically, by coupling the FET 421 in series between the ACS terminal 408 and the diode 410, the current that exits the ACS terminal 408 is limited to the saturation current of the FET 421.

In yet another exemplary embodiment, the simplified circuit shown in FIG. 4C can be modified to include an AC shorting capacitor placed in parallel with the diode 410. The simplified circuit of FIG. 4F can be used to compensate for certain undesirable nonlinearities present in a full circuit application. In some embodiments, due to parasitics present in the MOSFET layout, nonlinearity characteristics existing in the diode 410 of FIG. 4C may introduce undesirable nonlinearities in a full circuit implementation. As the diode is in place to provide DC bias conditions and is not intended to have any AC signals across it, it may be desirable in some embodiments to take steps to mitigate the effects of any AC signal present across the diode 410.

As shown in the simplified circuit of FIG. 4F, the circuit of FIG. 4C has been modified to include an AC shorting capacitor 423 wherein the AC shorting capacitor 423 is configured in parallel across the diode 410. The AC shorting capacitor 423 is placed in parallel with the diode 410 to ensure that nonlinearities of the diode 410 are not excited by an AC signal. In some exemplary circuits, such as in an RF switch, the AC shorting capacitor 423 does not impact the higher level full circuit, as the gate terminal 402 and the ACS terminal 408 typically have the same AC signal applied (i.e., AC equipotential).

In some circuit embodiments, body nodes of a multi-finger FET implementation may be connected to one another (using, for example, metal or silicon), overlapping the source fingers. On another side of the FET implementation, gate nodes may be are connected to one another (using, for example, metal or silicon) overlapping the drain fingers. As a result of this FET implementation, additional capacitance may result between the source and body (S−B), and further additional capacitance may result between the drain and gate (D−G). These additional capacitances may degrade the symmetry of the intrinsic device. Under AC excitation, this results in the gate terminal following the drain terminal more closely, and the body terminal following the source terminal more closely, which effectively creates an AC signal across the diode 410, which can excite nonlinearities of the diode 410 as described above. Using the exemplary embodiment shown in FIG. 4F, parasitic nonlinear excitation due to the overlapping fingers is mitigated.

Improved C_offPerformance Characteristics of ACC MOSFETs Made in Accordance with the Present Disclosed Method and Apparatus

FIG. 4G is a plot 460 of the off-state capacitance (C_off) versus an applied drain-to-source voltage of an SOI MOSFET when an AC signal is applied to the MOSFET. In one embodiment, a gate voltage equals −2.5 Volts+Vd/2, and Vs equals 0. A first plot 462 shows the off-state capacitance C_offof a typical prior art NMOSFET operating within the accumulated charge regime and thereby having an accumulated charge as described above with reference to FIG. 1. As shown in FIG. 4G, the off-state capacitance C_offshown in plot 462 of the prior art FET is voltage-dependent (i.e., it is nonlinear) and peaks when Vd=0 Volts. A second plot 464 illustrates the off-state capacitance C_offof an improved ACC SOI MOSFET made in accordance with the present teachings, wherein the accumulated charge is conveyed from the ACC MOSFET, thereby reducing, controlling and/or eliminating the accumulated charge from the ACC MOSFET body. As shown in FIG. 4G, the off-state capacitance C_offshown in plot 464 of the ACC SOI MOSFET is not voltage-dependent (i.e., it is linear).

As described above with reference to FIG. 2A, by controlling, reducing or eliminating the accumulated charge, the impedance 212 of the NMOSFET body 312 (FIG. 3C, and shown as the MOSFET body 114 in the electrical model of FIG. 2A) is increased to a very large value. This increase in the impedance 212 of the MOSFET body reduces the contribution to C_offcaused by the impedance of the junctions 218 and 220 (FIG. 2A), thereby reducing the overall magnitude of C_offand the nonlinear effects associated with the impedances of the

junctions

218 and 220. Plot 464 illustrates how the present teachings advantageously reduce both the nonlinearity and overall magnitude of the off-state capacitance C_offof the MOSFET. The reduced nonlinearity and magnitude of the off-state capacitance C_offimproves the performance of circuits using MOSFETs operating in an accumulated charge regime, such as RF switching circuits. Exemplary RF switching circuits implemented with the ACC MOSFETs described above with reference to FIGS. 4A-4F are now described with reference to FIGS. 5A-5D.

Exemplary Improved Performance RF Switch Implementations Using ACC SOI MOSFETs in Accordance with the Present Teachings

FIG. 5A shows a schematic diagram of a single pole, single throw (SPST) RF switch circuit 500. The RF switch circuit 500 is one example of a general class of well-known RF switch circuits. Similar RF switch circuits are described in the following co-pending and commonly assigned U.S. applications and patent: Provisional Application No. 60/651,736, filed Feb. 9, 2005, entitled “UNPOWERED SWITCH AND BLEEDER CIRCUIT;” application Ser. No. 10/922,135, filed Aug. 18, 2004, pending, which is a continuation application of application Ser. No. 10/267,531, filed Oct. 8, 2002, which issued Oct. 12, 2004 as U.S. Pat. No. 6,804,502, entitled “SWITCH CIRCUIT AND METHOD OF SWITCHING RADIO FREQUENCY SIGNALS”. Application Ser. No. 10/267,531, filed Oct. 8, 2002, which issued Oct. 12, 2004 as U.S. Pat. No. 6,804,502 claims the benefit of U.S. Provisional Application No. 60/328,353, filed Oct. 10, 2001. All of the above-cited patent applications and issued patent set forth above are hereby incorporated by reference herein as if set forth in full for their teachings on RF switch circuits including SOI MOSFET switch circuits.

Referring again to FIG. 5A, an SOI NMOSFET 506 is adapted to receive an RF input signal “RFin” at an input terminal 502. The SOI MOSFET 506 is electrically coupled to selectively couple the RFin input signal to an output terminal 504 (i.e., thereby convey an RF output signal Rfout at the output terminal 504). In the exemplary embodiment, the SOI NMOSFET 506 is controlled by a first control signal C1 that is conveyed by a control line 512 through a gate resistor 510 (optionally included for suppression of parasitic RF coupling). The control line 512 is electrically coupled to a control circuit 520, which generates the first control signal C1.

Referring again to FIG. 5A, an SOI NMOSFET 508 is adapted to receive the RF input signal RFin at its drain terminal, and to selectively shunt the input signal RFin to ground via an optional load resistor 518. The SOI NMOSFET 508 is controlled by a second control signal C1x which is conveyed by a control line 516 through a gate resistor 514 (optionally included for suppression of parasitic RF coupling and for purposes of voltage division). The control line 516 is electrically coupled to the control circuit 520, which generates the second control signal C1x.

The first and second control signals, C1 and C1x, respectively, are generated so that the SOI NMOSFET 506 operates in an on-state when the SOI NMOSFET 508 operates in an off-state, and vice versa. These control signals provide the gate bias voltages Vg to the gate terminals of the

NMOSFETs

506 and 508. When either of the

NMOSFETs

506 or 508 is biased to select the transistor off-state, the respective Vg must comprise a sufficiently large negative voltage so that the respective NMOSFET does not enter, or approach, an on-state due to the time varying applied voltages of the RF input signal RFin. The maximum power of the RF input signal RFin is thereby limited by the maximum magnitude of the gate bias voltage Vg (or, more generally, the gate-to-source operating voltage, Vgs) that the

SOI NMOSFETs

506 and 508 can reliably sustain. For RF switching circuits such as those exemplified herein, Vgs(max)=Vg+Vds(max)/2, where Vds=Vd−Vs, and Vds(max) comprises the maximum Vds due to the high-power input signal voltage levels associated with the RF input signal RFin.

Exemplary bias voltages for the

SOI NMOSFETs

506 and 508 may include the following: with V_thapproximately zero volts, Vg, for the on-state, of +2.5 V, and Vg, for the off-state, of −2.5 V. For these bias voltages, the SOI NMOSFETs may eventually operate in an accumulated charge regime when placed into their off-states. In particular, and as described above with reference to FIG. 2B, when the NMOSFET 506 is in the on-state, and the NMOSFET 508 is biased in the off-state, the output signal RFout may become distorted by the nonlinear behavior of the off capacitance C_offof the NMOSFET 508 caused by the accumulated charge. Advantageously, the improved ACC MOSFETs made in accordance with the present teachings can be used to improve circuit performance, especially as it is adversely affected by the accumulated charge.

FIG. 5B is a schematic of an improved RF circuit 501 adapted for higher performance using the present accumulated charge reduction and control techniques. The switch circuit 501 differs from the prior art circuit 500 (FIG. 5A) in that the NMOSFET 508 is replaced by an ACC NMOSFET 528 made in accordance with the present teachings. The ACC NMOSFET 528 is analogous to the ACC NMOSFET described above with reference to FIGS. 4A and 4B. Similarly, the gate, source, drain and ACC terminals of the ACC NMOSFET 528 are analogous to the respective terminals of the ACC NMOSFET 300. With the exception of the improved switch performance afforded by the improved ACC NMOSFET 528, the operation of the RF switch circuit 501 is very similar to the operation of the RF switch circuit 500 described above with reference to FIG. 5A.

Exemplary bias voltages for the NMOSFET 526 and the ACC NMOSFET 528 may include: with V_thapproximately zero, Vg, for the on-state, of +2.5 V, and Vg, for the off-state, of −2.5 V. For these bias voltages, the SOI NMOSFETs may operate in an accumulated charge regime when placed into the off-state. However, when the NMOSFET 526 is in the on-state and the ACC NMOSFET 528 is in the off-state, the output signal RFout at the output terminal 505 will not be distorted by nonlinear behavior of the off-state capacitance C_offof the improved ACC NMOSFET 528 due to the accumulated charge. When the ACC NMOSFET 528 operates in the accumulated charge regime, the accumulated charge is removed via the ACS terminal 508′. More specifically, because the gate terminal 502′ of the ACC NMOSFET 528 is connected to the ACS terminal 508′, the accumulated charge is removed or otherwise controlled as described above in reference to the simplified circuit of FIG. 4B. The control of the accumulated charge improves performance of the switch 501 by improving the linearity of the off transistor, 528, and thereby reducing the harmonic and intermodulation distortion of the RF output signal Rfout generated at the output terminal 505.

FIG. 5C is a schematic of another embodiment of an improved RF switch circuit 502 adapted for higher performance using the accumulated charge control techniques of the present disclosure. The switch circuit 502 differs from the prior art circuit 500 (FIG. 5A) in that the NMOSFET 508 is replaced by an ACC NMOSFET 528 made in accordance with the present teachings. The ACC NMOSFET 528 is analogous to the ACC NMOSFET 300 described above with reference to FIGS. 4A and 4C. Similarly, the gate, source, drain and ACC terminals of the ACC NMOSFET 528 are analogous to the respective terminals of the ACC NMOSFETs 300 described above with reference to FIGS. 4A and 4C. With the exception of the improved switch performance afforded by the improved ACC NMOSFET 528, the operation of the switch circuit 502 is very similar to the operations of the

switch circuits

500 and 501 described above with reference to FIGS. 5A and 5B, respectively.

Exemplary bias voltages for the NMOSFET 526 and the ACC NMOSFET 528 may include the following: with V_thapproximately zero volts, Vg, for the on-state, of +2.5 V, and Vg, for the off-state, of −2.5 V. For these bias voltages, the

SOI NMOSFETs

526, 528 may operate in an accumulated charge regime when placed into an off-state. However, when the NMOSFET 526 is in the on-state and the ACC NMOSFET 528 is in the off-state, the output signal RFout will not be distorted by nonlinear behavior of the off-state capacitance C_offof the ACC NMOSFET 528 due to the accumulated charge. Because the gate terminal 502′ of the ACC NMOSFET 528 is connected to the ACS terminal 508′ via a diode 509, the accumulated charge is entirely removed, reduced or otherwise controlled, as described above with reference to FIG. 4C. Similar to the improved switch 501 described above with reference to FIG. 5B, control of the accumulated charge improves performance of the switch 502 by improving the linearity of the off transistor, 528, and thereby reducing the harmonic and intermodulation distortion of the RF output signal Rfout output of the RF output terminal 505. Connection of the diode 509 as shown may be desired in some embodiments for suppression of positive current flow into the ACC NMOSFET 528 when it is biased into an on-state, as described above with reference to FIG. 4C.

FIG. 5D is a schematic of another embodiment of an improved RF switch circuit 503 adapted for higher performance using the present accumulated charge control techniques. The switch circuit 503 differs from the prior art circuit 500 (FIG. 5A) in that the NMOSFET 508 of FIG. 5A is replaced by an ACC NMOSFET 528 made in accordance with the present teachings. The ACC NMOSFET 528 is analogous to the ACC NMOSFET described above with reference to FIGS. 4A and 4D. With the exception of the improved switch performance afforded by the improved ACC NMOSFET 528, the operation of the switch circuit 503 is very similar to the operations of the

switch circuits

500, 501 and 502 described above with reference to FIGS. 5A-5C, respectively.

SOI NMOSFETs

526, 528 may operate in an accumulated charge regime when placed into the off-state. However, when the NMOSFET 526 is in the on-state and the ACC NMOSFET 528 is in the off-state, the output signal RFout produced at the output terminal 505 will not be distorted by the nonlinear behavior of the off-state capacitance C_offof the ACC NMOSFET 528 due to the accumulated charge. When the NMOSFET 528 operates in the accumulated charge regime, the accumulated charge is removed via the ACS terminal 508′. More specifically, because the ACS terminal 508′ of the ACC NMOSFET 528 is electrically coupled to the control circuit 520 via the control line 517 (i.e., controlled by the control signal “C2” as shown), the accumulated charge can be eliminated, reduced or otherwise controlled by applying selected bias voltages to the ACS terminal 508′ as described above with reference to FIG. 4D. Those skilled in the arts of electronic circuit design shall understand that a wide variety of bias voltage signals can be applied to the ACS terminal for the purpose of reducing or otherwise controlling the accumulated charge. The specific bias voltages may be adapted for use in a particular application. The control of the accumulated charge improves performance of the switch 503 by improving the linearity of the off-state transistor, 528, and thereby reducing the harmonic and intermodulation distortion of the RF output signal Rfout generated at the output terminal 505.

In the circuits described above with respect to FIGS. 5B-5D, the switching SOI MOSFETs 526 are shown and described as implemented using SOI MOSFETs of the prior art (i.e., they do not comprise ACC MOSFETs and therefore do not have an ACS terminal). Those skilled in the electronic device design arts shall understand and appreciate that in other embodiments of the disclosed method and apparatus, the prior art switching SOI MOSFETs 526 may be replaced, as required, by ACC SOI MOSFETs made in accordance with the present disclosure. For example, in some embodiments of RF switches implemented using the ACC MOSFET of the present disclosure, the RF switch comprises a single-pole double-throw RF switch. In this embodiment, the switching SOI MOSFETs (e.g., analogous to the switching ACC SOI MOSFETs 526 described above with reference to FIGS. 5B-5D) may comprise ACC SOI MOSFETs. Such an implementation prevents nonlinear behavior of the off-state switching SOI MOSFETs (which is turned off when it is not selected as an input “pole”) from detrimentally affecting the output of the RF signal as switched through the selected “pole”. This implementation is described in more detail below with reference to the RF switch circuit shown in FIG. 9. Many other examples will be apparent to those skilled in the arts of electronic circuits.

Exemplary RF Switch Implementation Using Stacked Transistors

In the exemplary embodiments of RF switch circuits described above, the switch circuits are implemented using a single SOI NMOSFET (e.g., the single SOI NMOSFET 506 of FIG. 5A, and the single SOI NMOSFET 526 of FIGS. 5B-5D) that selectively couples or blocks (i.e., electrically opens the circuit connection) the RF input signal to the RF output. Similarly, in the exemplary embodiments described above with reference to FIGS. 5A-5D, a single SOI NMOSFET (e.g., the single SOI NMOSFET 508 of FIG. 5A, and ACC SOI NMOSFET 528 of FIGS. 5B-5D) is used to shunt (FET in the on-state) or block (FET in the off-state) the RF input signal to ground. Commonly assigned U.S. Pat. No. 6,804,502, entitled “SWITCH CIRCUIT AND METHOD OF SWITCHING RADIO FREQUENCY SIGNALS”, issued Oct. 12, 2004, describes RF switch circuits using SOI NMOSFETs implemented with stacked transistor groupings that selectively couple and block RF signals.

One example of how stacked NMOSFETs may be implemented in accordance with the teachings of the present disclosure is illustrated in FIG. 6. An RF switch circuit 600 is analogous to the RF switch circuit 503 of FIG. 5D, wherein the single SOI NMOSFET 526 is replaced by a stack of

SOI NMOSFETs

602, 604 and 606. Similarly, the single ACC SOI NMOSFET 528 is replaced by a stack of

ACC SOI NMOSFETs

620, 622 and 624. The control signal C2 is provided to the ACS terminals of the

ACC SOI NMOSFETs

620, 622 and 624 via

optional resistors

626, 628, and 630, respectively. The

resistors

626, 628, and 630 may optionally be included in order to suppress parasitic RF signals between the stacked

ACC SOI NMOSFETs

620, 622, and 624, respectively. The RF switch circuit 600 operates analogously to the operation of the RF switch circuit 503 described above with reference to FIG. 5D.

Three stacked ACC SOI NMOSFETs are shown in each ACC NMOSFET stack in the exemplary stacked RF switch circuit 600 of FIG. 6. A plurality of three ACC NMOSFETs is shown for illustrative purposes only, however, those skilled in the integrated circuit design arts will understand that an arbitrary plurality may be employed according to particular circuit requirements such as power handling performance, switching speed, etc. A smaller or larger plurality of stacked ACC NMOSFETs may be included in a stack to achieve a desired operating performance.

Other stacked RF switch circuits, adapted for accumulated charge control, analogous to the circuits described above with reference to FIGS. 5B-5D, may also be employed. Implementations of such circuits shall be obvious from the teachings above to those skilled in the electronic device design arts, and therefore are not described further herein.

Exemplary Method of Operation

FIG. 7 illustrates an exemplary method 700 of improving the linearity of an SOI MOSFET having an accumulated charge sink (ACS) in accordance with the present disclosure. The method 700 begins at a STEP 702, whereat an ACC SOI MOSFET having an ACS terminal is configured to operate in a circuit. The ACS terminal may be operatively coupled to the gate of the SOI MOSFET (as described above with reference to FIGS. 4B, 4C, 5B and 5C), or to a control circuit (as described above with reference to FIGS. 4D and 5D). In other embodiments, the ACS terminal may be operatively coupled to any convenient accumulated charge sinking mechanism, circuit, or device as is convenient to the circuit or system designer. The method then proceeds to a step 704.

At the STEP 704, the ACC SOI MOSFET is controlled, at least part of the time, so that it operates in an accumulated charge regime. In most embodiments, as described above, the ACC MOSFET is operated in the accumulated charge regime by applying bias voltages that place the FET in an off-state condition. In one exemplary embodiment, the ACC SOI MOSFET comprises an ACC SOI NMOSFET that is configured as part of a shunting circuit of an RF switch. According to this exemplary embodiment, the SOI NMOSFET may be operated in an accumulated charge regime after the shunting circuit is placed into an off-state by applying a negative bias voltage to the gate terminal of the ACC NMOSFET.

The method then proceeds to a STEP 706, whereat the accumulated charge that has accumulated in the channel region of the ACC MOSFET is removed or otherwise controlled via the ACS terminal. In this embodiment, the accumulated charge is conveyed to another circuit terminal and is thereby reduced or otherwise controlled. One such exemplary circuit terminal that can be used to convey the accumulated charge from the MOSFET body comprises a gate terminal of the ACC MOSFET (see, e.g., the description above with reference to FIGS. 4B, 4C, 5B and 5C). Another exemplary circuit terminal that can be used to remove or otherwise control the accumulated charge comprises the terminal of a control circuit (see, e.g., FIGS. 4D and 5D). As described in more detail above, removing or otherwise controlling the accumulated charge in the ACC MOSFET body improves the linearity of the off-state ACC MOSFET, which reduces the harmonic distortion and IMD of signals affected by the ACC MOSFET, and which, in turn, improves circuit and system performance. In RF switch circuits, improvements (in both linearity and magnitude) are made to the off capacitance of shunting ACC MOSFET devices, which, in turn, improves the performance of the RF switch circuits. In addition to other switch performance characteristics, the harmonic and intermodulation distortions of the RF switch are reduced using the ACC method and apparatus of the present teachings.

FIGS. 8 and 9 show schematics of additional exemplary embodiments of RF switching circuits made in accordance with the disclosed method and apparatus for use in improving Linearity of MOSFETs having an ACS. As described in more detail below with reference to FIGS. 8 and 9, in some exemplary embodiments of RF switch circuits made in accordance with the present disclosure, it may be desirable to include drain-to-source resistors, R_ds, and thereby improve some switch performance characteristics when the switch is used in a particular application. These exemplary RF switch circuits are now described in more detail.

Exemplary RF Switch Implementations Using Stacked Transistors Having Source to Drain Resistors

FIG. 8 shows one exemplary embodiment of an RF switch circuit 800 made in accordance with the present disclosure. As shown in FIG. 8, some embodiments of RF switches made in accordance with the present disclosure may include drain-to-source (R_ds) resistors electrically connected to the respective sources and drains of the ACC MOSFETs. For example, the exemplary switch 800 of FIG. 8 includes drain-to-source R_dsresistors 802, 804, and 806 electrically connected to the respective sources and drains of the shunting

ACC SOI NMOSFETs

620, 622, and 624, respectively. Motivation for use of the drain-to-source R_dsresistors is now described.

As shall be appreciated by skilled persons from the present teachings, removal of the accumulated charge via the ACS terminal causes current to flow from the body of the ACC SOI MOSFET. For example, when a hole current flows from the body of an ACC SOI MOSFET via the ACS, an equal electron current flows to the FET source and/or drain. For some circuits (e.g., the RF switch circuit of FIG. 8), the sources and/or drains of the ACC SOI NMOSFETs are connected to other SOI NMOSFETs. Because off-state SOI NMOSFETs have a very high impedance (e.g., in the range of 1 Gohm for a 1 mm wide SOI NMOSFET), even a very small drain-to-source current (e.g., in the range of 1 nA) can result in an unacceptably large drain-to-source voltage Vds across the ACC SOI NMOSFET in satisfaction of Kirchhoff's well known current and voltage laws. In some embodiments, such as that shown in the RF switch circuits of FIGS. 8 and 9, such resultant very large drain-to-source voltages Vds undesirably impacts reliability and linearity of the ACC SOI NMOSFET. The drain-to-source resistors R_dsprovide a path between the ACC FET drain and source whereby currents associated with controlling the accumulated charge may be conducted away from the sources and drains of ACC SOI NMOSFETs when implemented in series with high impedance elements such as other ACC SOI NMOSFETs.

Exemplary operating voltages for the NMOSFETs 602-606 of FIG. 8, and the ACC NMOSFETs 620-624, may include the following: V_thapproximately zero volts, Vg, for the on-state, of +2.5 V, and Vg, for the off-state, of −2.5 V. In an exemplary embodiment, the ACC SOI NMOSFET 622 of FIG. 8 may have a width of 1 mm, and an electron-hole pair generation rate for accumulated charge producing a current of 10 pA/μm for operation in the accumulated charge regime. For the electron current supplied equally by the source and drain, and an impedance of the

ACC SOI NMOSFETs

620 and 622 on the order of 1 Gohm, then an unacceptable bias of −5 V would result on the source and drain of the ACC SOI NMOSFET 622 without the presence of R_dsresistors 802 and 806. This bias voltage would also be applied to the interior nodes of the

ACC SOI NMOSFETs

620 and 624.

Even currents smaller than the exemplary currents may produce adverse affects on the operation of the RF switching circuit 800 by reducing Vgs and/or Vgd of the ACC SOI MOSFETs 620-624 in the off-state, thereby reducing the power handling capability and reliability of the circuit by increasing leakage (e.g., when either Vgs or Vgd approaches V_th), by increasing hot-carrier damage caused by excess leakage, etc. Linearity of the MOSFETs is also degraded by reducing Vgs and/or Vgd when either value approaches V_th.

Exemplary values for the R_dsresistors 802 to 806 may be selected in some embodiments by selecting a value approximately equal to the resistance of the gate resistors 632-636 divided by the number of ACC SOI NMOSFETs in the stack (in the exemplary embodiment, there are three ACC FETs in the stack). More generally, the value of the R_dsresistors may be equal to the gate resistor value divided by the number of ACC SOI NMOSFETs in the stack. In one example, a stack of eight ACC SOI NMOSFETs may have gate resistors of 80 kohm and R_dsresistors of 10 kohm.

In some embodiments, the R_dsresistors may be selected so that they do not adversely affect switch performance characteristics, such as, for example, the insertion loss of the switch 800 due to the off-state ACC SOI NMOSFETs. For example, for a net shunt resistance greater than 10 kohm, the insertion loss is increased by less than 0.02 dB.

In other embodiments, the R_dsresistors may be implemented in circuits comprising a single ACC SOI MOSFET (as contrasted with the stacked shunting configuration exemplified in FIG. 8 by the shunting

ACC FETs

620, 622 and 624). For example, such circuits may be desirable if there are other high-impedance elements configured in series with an ACC SOI MOSFET that may cause a significant bias voltage to be applied to the source or drain as a result of the current flow created when removing or otherwise controlling accumulated charge. One exemplary embodiment of such a circuit is shown in FIG. 9.

FIG. 9 shows an exemplary single-pole double-throw (SPDT) RF switch circuit 900 made in accordance with the present teachings. As shown in FIG. 9, a DC blocking capacitor 904 is connected to a first RF input node 905 that receives a first RF input signal RF1. Similarly, a DC blocking capacitor 906 is connected to a second RF input node 907 that receives a second RF input signal RF2. Further, a DC blocking capacitor 902 is electrically connected to an RF common output node 903 that provides an RF common output signal (RFC) selectively conveyed to the node RFC 903 by the switch circuit 900 from either the first RF input node 905 or the second RF input node 907 (i.e., RFC either outputs RF1 or RF2, depending upon the operation of the switch as controlled by the control signals C1 and C1x described below in more detail).

A first control signal C1 is provided to control the operating states of the

ACC SOI NMOSFETs

526 and 528′ (i.e., C1 selectively operates the FETs in the on-state or the off-state). Similarly, a second control signal C1x is provided to control the operating states of the

ACC SOI NMOSFETs

528 and 526′. As is well known, and as described for example in the above incorporated commonly assigned U.S. Pat. No. 6,804,502, the control signals C1 and C1x are generated so that the

ACC SOI NMOSFETs

526 and 528′ are in an on-state when the

ACC SOI NMOSFETs

528 and 526′ are in an off-state, and vice versa. This configuration allows the RF switch circuit 900 to selectively convey either the signal RF1 or RF2 to the RF common output node 903.

A first ACS control signal C2 is configured to control the operation of the ACS terminals of the

SOI NMOSFETs

526 and 528′. A second ACS control signal C2x is configured to control the ACS terminals of the

ACC SOI NMOSFETs

528 and 526′. The first and second ACS control signals, C2 and C2x, respectively, are selected so that the ACSs of the associated and respective NMOSFETs are appropriately biased in order to eliminate, reduce, or otherwise control their accumulated charge when the ACC SOI NMOSFETs operate in an accumulated charge regime.

As shown in the RF switch circuit 900 of FIG. 9, in some embodiments, an R_dsresistor 908 is electrically connected between the source and drain of the switching ACC NMOSFET 526. Similarly, in some embodiments, an R_dsresistor 910 is electrically connected between the source and drain of the switching ACC NMOSFET 526′. According to this example, the circuit 900 is operated so that either the shunting ACC NMOSFET 528 or the shunting ACC NMOSFET 528′ operate in an on-state at any time (i.e., at least one of the input signals RF1 at the node 905 or RF2 at the node 907 is always conveyed to the RFC node 903), thereby providing a low-impedance path to ground for the

node

905 or 907, respectively. Consequently, either the R_dsresistor 908 or the R_dsresistor 910 provides a low-impedance path to ground from the RF common node 903, thereby preventing voltage bias problems caused as a result of ACC current flow into the

nodes

903, 905 and 907 that might otherwise be caused when using the

DC blocking capacitors

902, 904 and 906.

Additional Exemplary Benefits Afforded by the ACC MOSFETs of the Present Disclosure

As described above, presence of the accumulated charge in the bodies of the SOI MOSFETs can adversely affect the drain-to-source breakdown voltage (BVDSS) performance characteristics of the floating body MOSFETs. This also has the undesirable affect of worsening the linearity of off-state MOSFETs when used in certain circuits such as RF switching circuits. For example, consider the shunting SOI NMOSFET 528 shown in FIG. 9. Further consider the case wherein the shunting NMOSFET 528 is implemented with a prior art SOI NMOSFET, rather than with the ACC NMOSFET made in accordance with the present teachings. Assume that the RF transmission line uses a 50-ohm system. With small signal inputs, and when the NMOSFET 528 operates in an off-state, the prior art off-state shunting NMOSFET 528 may introduce harmonic distortion and/or intermodulation distortion in the presence of multiple RF signals This will also introduce a noticeable loss of signal power.

When sufficiently large signals are input that cause the NMOSFET 528 to enter a BVDSS regime, some of the RF current is clipped, or redirected through the NMOSFET 528 to ground, resulting in a loss of signal power. This current “clipping” causes compression behavior that can be shown, for instance, in a RF switch “Pout vs. Pin” plot. This is frequently characterized by P1 dB, wherein the insertion loss is increased by 1.0 dB over the small-signal insertion loss. This is an obvious indication of nonlinearity of the switch. In accordance with the present disclosed method and apparatus, removing, reducing or otherwise controlling the accumulated charge increases the BVDSS point. Increases to the BVDSS point of the NMOSFET 528 commensurately increases the large-signal power handling of the switch. As an example, for a switch, doubling the BVDSS voltage of the ACC NMOSFET increases the P1 dB point by 6 dB. This is a significant accomplishment as compared with the prior art RF switch designs.

In addition, as described above in more detail, presence of the accumulated charge in SOI MOSFET body adversely impacts the magnitude of Coff and also takes time to form when the FET is switched from an on-state to an off-state. In terms of switch performance, the nonlinearity of C_offadversely impacts the overall switch linearity performance (as described above), and the magnitude of C_offadversely affects the small-signal performance parameters such as insertion loss, insertion phase (or delay), and isolation. By reducing the magnitude of C_offusing the present disclosed method and apparatus, the switch (implemented with ACC MOSFETs) has reduced insertion loss due to lowered parasitic capacitance, reduced insertion phase (or delay), again due to lowered parasitic capacitance, and increased isolation due to less capacitive feedthrough.

The ACC MOSFET also improves the drift characteristic of SOI MOSFETs as pertains to the drift of the small-signal parameters over a period of time. As the SOI MOSFET takes some time to accumulate the accumulated charge when the switch is off, the C_offcapacitance is initially fairly small. However, over a period of time while operated in the accumulated charge regime, the off-state capacitance C_offincreases toward a final value. The time it takes for the NMOSFET to reach a full accumulated charge state depends on the electron-hole pair (EHP) generation mechanism. Typically, this time period is on the order of approximately hundreds of milliseconds for thermal EHP generation at room temperature, for example. During this charge-up time period, the insertion loss and insertion phase increase. Also, during this time period, the isolation decreases. As is well known, these are undesirable phenomena in standard SOI MOSFET devices. These problems are alleviated or otherwise mitigated using the ACC NMOSFETs and related circuits described above.

In addition to the above-described benefits afforded by the disclosed ACC MOSFET method and apparatus, the disclosed techniques also allow the implementation of SOI MOSFETs having improved temperature performance, improved sensitivity to Vdd variations, and improved sensitivity to process variations. Other improvements to the prior art SOI MOSFETs afforded by the present disclosed method and apparatus will be understood and appreciated by those skilled in the electronic device design and manufacturing arts.

Exemplary Fabrication Methods

In one embodiment of the present disclosure, the exemplary RF switches described above may be implemented using a fully insulating substrate silicon-on-insulator (SOI) technology. Also, as noted above, in addition to the commonly used silicon-based systems, some embodiments of the present disclosure may be implemented using silicon-germanium (SiGe), wherein the SiGe is used equivalently in place of silicon.

In some exemplary embodiments, the MOSFET transistors of the present disclosure may be implemented using “UltraThin-Silicon (UTSi)” (also referred to herein as “ultrathin silicon-on-sapphire”) technology. In accordance with UTSi manufacturing methods, the transistors used to implement the inventive methods disclosed herein are formed in an extremely thin layer of silicon in an insulating sapphire wafer. The fully insulating sapphire substrate enhances the performance characteristics of the inventive RF circuits by reducing the deleterious substrate coupling effects associated with non-insulating and partially insulating substrates. For example, insertion loss improvements may be realized by lowering the transistor on-state resistances and by reducing parasitic substrate conductance and capacitance. In addition, switch isolation is improved using the fully insulating substrates provided by UTSi technology. Owing to the fully insulating nature of silicon-on-sapphire technology, the parasitic capacitance between the nodes of the RF switches is greatly reduced as compared with bulk CMOS and other traditional integrated circuit manufacturing technologies.

Examples of and methods for making silicon-on-sapphire devices that can be implemented in the MOSFETs and circuits described herein, are described in U.S. Pat. Nos. 5,416,043 (“Minimum charge FET fabricated on an ultrathin silicon on sapphire wafer”); 5,492,857 (“High-frequency wireless communication system on a single ultrathin silicon on sapphire chip”); 5,572,040 (“High-frequency wireless communication system on a single ultrathin silicon on sapphire chip”); 5,596,205 (“High-frequency wireless communication system on a single ultrathin silicon on sapphire chip”); 5,600,169 (“Minimum charge FET fabricated on an ultrathin silicon on sapphire wafer”); 5,663,570 (“High-frequency wireless communication system on a single ultrathin silicon on sapphire chip”); 5,861,336 (“High-frequency wireless communication system on a single ultrathin silicon on sapphire chip”); 5,863,823 (“Self-aligned edge control in silicon on insulator”); 5,883,396 (“High-frequency wireless communication system on a single ultrathin silicon on sapphire chip”); 5,895,957 (“Minimum charge FET fabricated on an ultrathin silicon on sapphire wafer”); 5,920,233 (“Phase locked loop including a sampling circuit for reducing spurious side bands”); 5,930,638 (“Method of making a low parasitic resistor on ultrathin silicon on insulator”); 5,973,363 (“CMOS circuitry with shortened P-channel length on ultrathin silicon on insulator”); 5,973,382 (“Capacitor on ultrathin semiconductor on insulator”); and 6,057,555 (“High-frequency wireless communication system on a single ultrathin silicon on sapphire chip”). All of these referenced patents are incorporated herein in their entirety for their teachings on ultrathin silicon-on-sapphire integrated circuit design and fabrication.

Similarly to other bulk and SOI CMOS processes, an SOS enhancement mode NMOSFET, suitable for some embodiments of the present disclosure, may, in some embodiments, be fabricated with a p-type implant into the channel region with n-type source and drain regions, and may have a threshold voltage of approximately +500 mV. The threshold voltage is directly related to the p-type doping level, with higher doping resulting in higher thresholds. Similarly, the SOS enhancement mode PMOSFET may, in some exemplary embodiments, be implemented with a n-type channel region and p-type source and drain regions. Again, the doping level defines the threshold voltage with higher doping resulting in a more negative threshold.

In some exemplary embodiments, an SOS depletion-mode NMOSFET, suitable for some embodiments of the present disclosure, may be fabricated by applying the p-type channel-implant mask to the n-type transistor, resulting in a structure that has n-type channel, source, and drain regions and a negative threshold voltage of approximately −500 mV. Similarly, in some exemplary embodiments, a suitable depletion-mode PMOSFET may be implemented by applying the n-type channel-implant mask to the p-type transistor, resulting in a structure that has p-type channel, source, and drain regions and a positive threshold voltage of approximately +500 mV.

A reference relating to the fabrication of enhancement-mode and depletion-mode transistors in SOS is “CMOS/SOS/LSI Switching Regulator Control Device,” Orndorff, R. and Butcher, D., Solid-State Circuits Conference, Digest of Technical Papers, 1978 IEEE International, Volume XXI, pp. 234-235, February 1978. The “Orndorff” reference is hereby incorporated in its entirety herein for its techniques on the fabrication of enhancement-mode and depletion-mode SOS transistors.

Embodiments of Methods and Apparatuses for Improving Gate Oxide Reliability in Accordance with the Present CIP

The present CIP describes methods and devices for improving gate oxide reliability of SOI MOSFETs using ACC techniques to control accumulated charge and the adverse effects thereof. Persons skilled in the arts of electronic devices will appreciate that the teachings herein apply equally to NMOSFETs and PMOSFETs. For simplicity, the embodiments and examples presented herein for illustrative purposes include only NMOSFETs, unless otherwise noted. By making well known changes to dopants, charge carriers, polarity of bias voltages, etc., persons skilled in the arts of electronic devices will easily understand how these embodiments and examples may be adapted for use with PMOSFETs.

Impact of Accumulated Charge on Gate Oxide Reliability

Referring to FIG. 1, the inventors have observed that the accumulated charge 120 has an adverse effect on the reliability of the gate oxide 110. This is a novel observation not known according to prior art. The adverse effects are caused by electric field lines from the gate 108 (due to the bias voltages Vg, Vs and Vd) primarily terminating at the accumulated charge 120. If the accumulated charge 120 is not present, the electric field lines from the gate 108 terminate at the boundaries of the undepleted regions of the source 112 and the drain 116. The

depletion regions

122 and 124 move the boundaries of the undepleted regions in the source 112 and the drain 116 away from the gate oxide. Thus, when the accumulated charge 120 is present the electric field stress on the gate oxide is much larger than when the accumulated charge 120 is prevented from forming, removed, reduced, or otherwise controlled.

Reduction or otherwise control of the accumulated charge therefore enables use of larger bias voltages for a given thickness of the gate oxide 110. Alternatively, reduction or control of the accumulated charge enables use of reduced gate oxide 110 thickness for given bias voltages. A combination of larger bias voltages and reduced gate oxide 110 thickness is also enabled by control of the accumulated charge. Larger bias voltages allow larger input voltages in the transition off-state, thereby improving power handling capability. Reduced gate oxide thickness provides improved insertion loss, thereby allowing either improvements in power handling capability, or an option of using smaller SOI NMOSFETs to control a given power level.

Although many of the examples herein relate to RF switches, persons skilled in the arts of electronic circuits will understand that the present teachings may also be applied to RF mixers, power amplifiers, level shifting circuits, negative voltage generators, oscillators, DC-DC converters and other circuits that employ SOI MOSFETs. In particular, persons skilled in the arts of RF circuits will readily understand how the present teachings may be implemented for RF power amplifiers such as described in the following U.S. patent applications: “STACKED TRANSISTOR METHOD AND APPARATUS,” application Ser. No. 10/875,405, filed Jun. 23, 2004, pending; and “INTEGRATED RF FRONT END,” application Ser. No. 11/158,597 (CIP of application Ser. No. 10/875,405, filed Jun. 23, 2004), filed Jun. 22, 2005, pending. These cited applications are commonly owned by the assignee of the present application, and are hereby fully incorporated by reference herein, as though set forth in full for teachings on using SOI MOSFETs for RF power amplifiers.

Accumulated Charge Control Using Bias Voltage Pulses

In one embodiment, applying bias voltage pulses comprises an ACC technique that may be used to control the accumulated charge 120. Referring again to FIG. 1, as described above, the accumulated charge 120 results from a slow electron-hole pair generation process that occurs when the gate bias voltage Vg is negative with respect to the source bias voltage Vs and the drain bias voltage Vd. If a positive voltage pulse above V_this applied to the gate terminal 104, an conducting channel comprising electrons is formed in the body 114 proximate the gate oxide 110, and the accumulated charge 120 is dissipated due to drift and recombination. When the gate bias voltage Vg returns to the negative bias level present prior to the application of the positive voltage pulse, the accumulated charge 120 regenerates in a time period having a time scale that is typically in the millisecond range or longer. Consequently, the accumulated charge 120 may be controlled by applying a series of positive voltage pulses to the gate terminal 104, repeated at a rate that is sufficient to prevent the accumulated charge 120 from accumulating.

Accumulated Charge Control Using RF Signal Voltage

In one embodiment, an applied RF signal voltage comprises an ACC technique. According to this embodiment, an NMOSFET is operated with a large series resistor electrically connected to the gate in series with the gate bias source (as described below in reference to FIG. 5A). The gate resistor is sufficiently large that the AC gate voltage tracks the AC drain voltage signal at one-half the amplitude of the quickly-varying drain voltage. This occurs because the gate is coupled to the drain and source by parasitic capacitors (e.g., by

capacitors

202 and 204 of FIG. 2A). Thus, a high-frequency drain voltage signal may be coupled to the gate, and the resulting gate voltage further coupled to the source, thereby dividing the voltage on the gate by a factor of two, providing that the gate resistance is sufficiently large that it does not shunt or attenuate the resulting high-frequency AC voltage induced on the gate. For example, if an applied gate bias of −2.5 V is placed on a switch gate with large series resistor, the gate voltage will vary around the −2.5 V bias level in response to an applied AC drain signal. If a 2 V amplitude high-frequency signal (i.e., a 4 V peak-to-peak RF signal) is applied to the drain as Vd, the gate voltage will be moved from the quiescent level of −2.5 V to −1.5 V at a maximum and to −3.5 V at a minimum. When the gate voltage moves to −1.5 V, the accumulated charge moves toward the level that would be present in equilibrium at −1.5 V, through diffusion to the source and drain, where the accumulated holes recombine with the large concentrations of electrons. Alternatively, the accumulated charge may recombine by other processes with other electron sources. When the gate voltage moves more negative, the accumulated charge will increase, but the time spent at the more negative voltage is too short for the accumulated charge to increase significantly before the voltage moves less negative again. I.e., the recombination process at the less negative voltage will occur more rapidly than at more negative voltages, while the generation process is largely independent of voltage. Consequently, under the dynamic conditions of the present example, the net effect of the AC voltage swing in the gate voltage is to reduce the effective level of accumulated charge relative to what it would be absent the AC signal. After many cycles of RF swing, a steady-state accumulated charge level will be present that is significantly less than would be present at the DC gate bias voltage of −2.5 V. According to the present example the effective accumulated charge may be approximately at the level that would be present for an effective gate bias of −1.5 V.

RF Switch Circuits Adapted for Improved Oxide Reliability Using Bias Voltage Pulses

FIG. 5A shows a schematic diagram of a single pole, single throw (SPST) RF switch circuit 500 that may be adapted to control accumulated charge and improve gate oxide reliability, in accordance with the present CIP. In particular, as described below, the RF switch circuit 500 may be operated using bias voltage pulse techniques to reduce or otherwise control accumulated charge. The RF switch circuit 500 is one example of a general class of well-known RF switch circuits. Similar RF switch circuits are described in the following co-pending and commonly assigned U.S. applications and patent: Provisional Application No. 60/651,736, filed Feb. 9, 2005, entitled “UNPOWERED SWITCH AND BLEEDER CIRCUIT;” application Ser. No. 10/922,135, filed Aug. 18, 2004, pending, which is a continuation application of application Ser. No. 10/267,531, filed Oct. 8, 2002, which issued Oct. 12, 2004 as U.S. Pat. No. 6,804,502, entitled “SWITCH CIRCUIT AND METHOD OF SWITCHING RADIO FREQUENCY SIGNALS”. Application Ser. No. 10/267,531, filed Oct. 8, 2002, which issued Oct. 12, 2004 as U.S. Pat. No. 6,804,502 claims the benefit of U.S. Provisional Application No. 60/328,353, filed Oct. 10, 2001. All of the above-cited applications and issued patent set forth above are hereby incorporated by reference herein as if set forth in full for their teachings on RF switch circuits including SOI MOSFET switch circuits.

NMOSFETs

506 and 508. When either of the

NMOSFETs

SOI NMOSFETs

The power that can be accommodated by the

SOI NMOSFETs

506, 508 is limited by insertion loss. Insertion loss can be improved by reducing gate oxide thicknesses. Therefore, as described above, the power handling performance of the

SOI NMOSFETs

506, 508 can be improved using the ACC techniques of the present disclosure which allow implementation of SOI MOSFETs having thinner gate oxides. Persons skilled in the electronic device and circuit design arts shall appreciate that an appropriate combination of higher bias voltages and reduced gate oxide thicknesses can be used to improve circuit performance.

Exemplary bias voltages for the

SOI NMOSFETs

506 and 508 may include the following: with V_thapproximately zero volts, Vg, for the on-state, of +2.5 V, and Vg, for the off-state, of −2.5 V. For these bias voltages, the SOI NMOSFETs may eventually operate in an accumulated charge regime when placed into their off-states. As described above in reference to FIG. 1, the maximum magnitude for the gate bias voltage Vg (or more generally, Vgs) that can be reliably sustained can be significantly improved by using the accumulated charge control techniques of the present disclosure.

Accumulated charge control may be implemented by pulsing bias voltages applied to the gates of the

SOI NMOSFETs

506, 508. In reference to FIG. 5A, the control signals C1 and C1x may be selectively controlled by the control circuit 520 to provide such bias voltage pulses to the gates of the

SOI NMOSFETs

506, 508. Any suitable method and components can be used to provide the bias voltage pulses. In an exemplary embodiment, a pulse generation circuit is used to generate all signals required to control the pulse process.

The gate bias voltage pulses are provided at any suitable magnitude, interval, and duration in order to reduce accumulated charge to the desired level. It is to be understood that each of these variables is dependent on the others and can be altered according to the application and preference of IC designer. For example, an individual IC may be capable of generating different pulses at different time periods and depending on any of a number of predetermined factors. By different, it is meant that the pulses may be of different magnitudes, intervals, durations, or a combination thereof. While a number of external factors can influence the nature of applied pulses according to the disclosure, recognize that one factor relevant thereto is the operating temperature of the SOI NMOSFET. At higher operating temperatures, accumulated charge builds up more rapidly. Thus, more frequent pulsing may be desirable under such conditions.

In one embodiment, a gate bias voltage pulse is applied to an off-state FET to force the FET toward an on-state. Although this gate bias voltage pulse need not exceed the threshold voltage (V_th) of the FET, exceeding V_this an exemplary embodiment of the present disclosure. The magnitude of the gate bias voltage pulse may be fixed or variable. For simplicity, a pulse of a fixed magnitude, to exceed V_thby a selected value (e.g., the pulse voltage may have a maximum value of V_th+0.5 V) may be used in an exemplary embodiment.

The magnitude of the pulsed gate voltage impacts the pulse duration needed to achieve the desired effects and vice versa. In one embodiment, the FET is pulsed for a duration of approximately 5 μs or less. This meets the GSM standard switching time requirements and further provides for preferred switching times of from 1 μs to 5 μs.

At the end of a gate bias voltage pulse, the FET is maintained in an off-state by the control signal. The time period for which the FET is maintained in the off-state varies depending on the application and on the pulsed gate bias voltage applied during the last pulse. In order to control the electric field stress in the gate oxide, it is desirable to apply the gate bias pulses at a sufficient frequency to prevent a build-up of accumulated charge proximate to the gate oxide. For example, if it is known that the accumulated charge build-up time is on the order of 100 mS or more, a repetition interval for the gate bias voltage pulses may be selected to be on the order of 10 mS or less.

RF Switch Circuits Adapted for Improved Oxide Reliability Using RF Signal Voltage

In another embodiment, an RF signal voltage may be used as an ACC technique. When the

SOI NMOSFETs

506 and 508 are in the off-state with the exemplary applied gate bias of −2.5 V, the gate voltage will vary around the −2.5 V bias level in response to an applied AC drain signal because the resistors 510 and 514 (these resistors may have exemplary values of 10 kohm) allow the gate voltages to track the AC drain signal at one-half amplitude as described above. If a 2 V amplitude high-frequency signal (i.e., a 4 V peak-to-peak RF signal) is applied to the drain of one of the

SOI NMOSFETs

506 or 508 in the off-state as Vd, the gate voltage will be moved from the quiescent level of −2.5 V to −1.5 V at a maximum and to −3.5 V at a minimum. When the gate voltage moves to −1.5 V, the accumulated charge moves toward the level that would be present in equilibrium at −1.5 V, through diffusion to the source and drain, where the accumulated holes recombine with the large concentrations of electrons. Alternatively, the accumulated charge may recombine by other processes with other electron sources. When the gate voltage moves more negative, the accumulated charge will increase, but the time spent at the more negative voltage is too short for the accumulated charge to increase significantly before the voltage moves less negative again. I.e., the recombination process at the less negative voltage will occur more rapidly than at more negative voltages, while the generation process is largely independent of voltage. Consequently, under the dynamic conditions of the present example, the net effect of the AC voltage swing in the gate voltage is to reduce the effective level of accumulated charge relative to what it would be absent the AC signal. After many cycles of RF swing, a steady-state accumulated charge level will be present that is significantly less than would be present at the DC gate bias voltage of −2.5 V. According to the present example the effective accumulated charge may be approximately at the level that would be present for an effective gate bias of −1.5 V.

Circuits Adapted for Improved Oxide Reliability Using ACC MOSFETs

In one embodiment, circuits having ACC SOI MOSFETs may be adapted for improved oxide reliability in accordance with the present CIP. The use of ACC SOI MOSFETs to control accumulated charge in RF switch circuits is described above in reference to FIGS. 5B-5D and FIG. 6, FIGS. 8, and FIG. 9. The ACC techniques as described in reference to these figures may also be adapted to improve gate oxide reliability, as described below in reference to FIG. 10. In other embodiments, the exemplary circuit including the ACC SOI MOSFET may include, without limitation, an RF mixer, a power amplifier, a level shifting circuit, a negative voltage generator, an oscillator, a DC-DC converter or other circuit using SOI MOSFETs.

Methods for Using ACC Techniques to Improve Gate Oxide Reliability

FIG. 10A is a flow chart of a method 1000 of implementing SOI MOSFETs using the ACC techniques of the present disclosure. At a STEP 1002, a first circuit including one or more SOI NMOSFETs is implemented to perform circuit functions according to the prior art (i.e., without ACC).

At a STEP 1004, the reliability of the gate oxides of the SOI NMOSFETs is determined with and without ACC. In one embodiment, the gate oxide reliability is determined by performing time dependent dielectric breakdown (TDDB) measurements on test SOI NMOSFETs corresponding to the SOI NMOSFETs included in the first circuit. The TDDB measurements provide information relating to the lifetime of the SOI NMOSFETs due to gate oxide breakdown. The results of the TDDB measurements indicate a maximum electric field, “E_tb”, in the oxide that can be sustained and yet still achieve a desired lifetime (i.e., an average “time-to-breakdown” or “tb”) for the SOI NMOSFETs. With a given value for E_tb, minimum gate oxide thicknesses and maximum bias voltages required to achieve the desired lifetime can be determined, as described below. The results of the TDDB measurements may also indicate a maximum gate-to-source bias voltage V_gs, that can be sustained and yet still achieve a desired lifetime for the SOI NMOSFETs. TDDB measurements are well known to persons skilled in the electronic device arts. Examples of TDDB measurement techniques are described in the above-incorporated paper by C. Hu and Q. Lu, the reference by Suehle and Chaparala, and reference by R. Bolam.

Reliability of the gate oxides is further determined with and without ACC by performing TDDB measurements of the SOI NMOSFETs operating with selected bias voltages. The measurement results provide information relating to the improved TDDB present in the gate oxides for selected bias conditions and gate oxide thicknesses. In one example of a TDDB measurement of an SOI NMOSFET, using a first gate bias V_g1=−10 V, Vs=0 V, Vd=0 V, and a gate oxide thickness T_ox1of 8.2 nm, an average time to failure of 500 seconds was determined without ACC. With ACC, the same average time to failure of 500 seconds required V_g2of −14V. Thus, according to this example, using ACC enables using a second V_g2that is increased by a factor of approximately 40% over V_g1, while maintaining an equivalent TDDB lifetime. Alternatively, using ACC enables using a second T_ox2that is equal to T_ox1multiplied by a factor of approximately 1/1.4, without an adverse effect on TDDB lifetime.

Referring again to the method 1000 of FIG. 10A, at a STEP 1006, a second circuit, performing the same function as the first circuit, is implemented using ACC, responsive (as described in more detail below) to the determination of the reliability of the gate oxides obtained at the previous STEP 1004. In one embodiment, the second circuit may include an ACC SOI NMOSFET having an ACS, operatively coupled to elements within the second circuit according to the teachings above. In another embodiment, the ACC technique may include use of bias voltage pulse techniques, as described above in reference to FIG. 5A. The SOI NMOSFETs in the second circuit may incorporate a reduced thickness for their gate oxides, smaller body widths, or any combination thereof. Further, the second circuit may also be configured to include a lesser plurality of SOI NMOSFETs. For example, the plurality of SOT NMOSFETs included in a stacked RF switch circuit in the second circuit may be less than the plurality of SOI MOSFETs in the corresponding stack of the first circuit.

At a STEP 1008, the second circuit is operated using the present ACC techniques, as described above. In one embodiment, the second circuit may comprise an RF switch operated at higher RF power levels than the first circuit.

FIG. 10B is a flow chart of another embodiment of a method 1001 of implementing SOI MOSFETs using the ACC techniques of the present disclosure. At a STEP 1012, a maximum electric field in the oxide, “E_ox1,”, is determined using well known simulation techniques (e.g., using “Medici” simulation techniques, etc.) for an SOT NMOSFET having a given oxide thickness Tox, operated without ACC. Bias voltages for Vg, Vs and Vd are selected to correspond to a circuit implementation including the SOI NMOSFET operating in the off-state. In one example, set forth above, the following values may be selected: Vg=−5 V; Vs=0 V; Vd=0 V; and a gate oxide thickness Tox of 8.2 nm. Using these selected values, for example, a maximum electric field E_ox1=4.3 MV/cm is determined in the gate oxide for an uncontrolled accumulated charge.

Referring to the method 1001 of FIG. 10B, at a STEP 1014, a maximum electric field in the oxide, “E_ox2”, is determined for the SOT NMOSFET, operated with ACC, using the same selected bias voltages used in accordance with the STEP 1012. For the illustrated example, for an accumulated charge controlled to zero value, a maximum electric field “E_ox2” of 2.1 MV/cm is determined. An improvement factor “F1”, is defined according to the following expression: F1=E_ox1/E_ox2.

At a STEP 1016, a circuit including the SOI NMOSFET is implemented responsive to the improvement factor F1. In one embodiment of the disclosed method, the circuit is implemented by selecting a maximum gate-to-source operating voltage Vgs for the off-state and Tox in accordance with the following equation:
Vgs(max)/Tox=F1·E_tb; (1)
wherein E_tbcomprises the maximum oxide electric field for which a given TDDB lifetime of the SOI NMOSFET can be obtained (e.g., ˜5 MV/cm for a lifetime of 10 years is a “rule of thumb” value). More generally, the SOI NMOSFET bias voltages, maximum signal voltages, and gate oxide thicknesses can be selectively improved using the ACC techniques of the present teachings. These improvements over the prior art SOI NMOSFET devices are provided according to the improvement factor F1, as shall be readily appreciated by persons skilled in the electronic device design arts. Finally, at a STEP 1018, the circuit is operated using ACC techniques as described above.

FIG. 10C is a flow chart of another embodiment of a method 1003 of implementing SOI MOSFETs using the present ACC techniques. At a STEP 1022, a limiting gate bias voltage, “V_g1”, is determined using TDDB measurements for an SOI NMOSFET having a given oxide thickness, Tox, operated in the off-state without ACC. For example, for Vd=Vs=0, the limiting gate bias voltage V_g1represents the most negative gate bias that can be sustained to obtain a desired operating lifetime without ACC.

At a STEP 1024, a limiting gate bias voltage V_g2is determined using TDDB measurements for the SOI NMOSFET having an oxide thickness Tox, operated in the off-state with ACC. For example, for Vd=Vs=0, the limiting gate bias voltage V_g2represents the most negative gate bias that can be sustained to obtain a desired operating lifetime with ACC. An improvement factor F2 is defined according to the following expression: F2=V_g1/V_g2.

At a STEP 1026, a circuit including the SOI NMOSFET is implemented responsive to the improvement factor, F2. In one embodiment of the inventive method, the circuit is implemented by selecting a maximum gate-to-source operating voltage Vgs for the off-state and Tox according to the following equation:
Vgs(max)/Tox=F2·E_tb (2)
More generally, the SOI NMOSFET bias voltages, maximum signal voltages, and gate oxide thicknesses can be selectively improved over the prior art according to the improvement factor F2, as shall be readily appreciated by persons skilled in the electronic device arts. Referring again to FIG. 10C, at a STEP 1028, as described above, the method terminates by operating the circuit using the disclosed ACC techniques.

A number of embodiments of the present inventive concept have been described. Nevertheless, it will be understood that various modifications may be made without departing from the scope of the inventive teachings. For example, it should be understood that the functions described as being part of one module may in general be performed equivalently in another module.

Accordingly, it is to be understood that the concepts described herein are not to be limited by the specific illustrated embodiments, but only by the scope of the appended claims.

Claims

What is claimed is:

1. An RF switch, comprising:

a first RF port;

a second RF port;

a first switch transistor grouping coupled with the first and second RF ports, and controlled by a first switch control signal, the first switch transistor grouping comprising a first plurality of switch NMOSFETs arranged in a stacked configuration;

a shunt transistor grouping coupled with the first RF port and with ground, and controlled by a shunt control signal, the shunt transistor grouping comprising a plurality of shunt NMOSFETs arranged in a stacked configuration, wherein at least one of the plurality of shunt NMOSFETs comprises a first body and a first accumulated charge sink (ACS) coupled with the first body and configured so that when the at least one shunt NMOSFET is disabled, a first negative bias voltage that is substantially negative with respect to ground applied to the first ACS substantially prevents accumulated charge from accumulating in the first body of the at least one shunt NMOSFET; and

a silicon-on-insulator substrate, wherein the first switch transistor grouping and the shunt transistor grouping are fabricated in a silicon layer of the silicon-on-insulator substrate so that when the first switch transistor grouping is enabled by the first switch control signal and the shunt transistor grouping is disabled by the shunt control signal, a signal on the first RF port is passed through to the second RF port, and when the first switch transistor grouping is disabled by the first switch control signal and the shunt transistor grouping is enabled by the shunt control signal, the signal on the first RF port is grounded.

2. The RF switch of claim 1, wherein at least one of the first plurality of switch NMOSFETs has a second body and a second accumulated charge sink (ACS) coupled with the second body and configured so that when the at least one switch NMOSFET is disabled, a second negative bias voltage that is substantially negative with respect to ground applied to the second ACS substantially prevents accumulated charge from accumulating in the second body of the at least one switch NMOSFET.

3. The RF switch of claim 2, wherein the at least one switch NMOSFET includes a gate, and wherein the gate of the at least one switch NMOSFET and the second ACS are coupled together and also coupled with the first switch control signal.

4. The RF switch of claim 2, wherein the at least one shunt NMOSFET includes a gate, and wherein the gate of the at least one shunt NMOSFET and the first ACS are coupled together and also coupled with the shunt control signal.

5. The RF switch of claim 2, wherein the at least one switch NMOSFET includes a gate, and wherein a diode is coupled between the at least one switch NMOSFET gate and the second ACS such that the diode prevents current flow into the second body when the at least one switch NMOSFET is enabled.

6. The RF switch of claim 2, wherein the at least one shunt NMOSFET includes a gate, and wherein a diode is coupled between the at least one shunt NMOSFET gate and the first ACS such that the diode prevents current flow into the first body when the at least one shunt NMOSFET is enabled.

7. The RF switch of claim 2, wherein the at least one switch NMOSFET includes a gate, and wherein an electrical device is coupled between the gate and the second ACS such that the electrical device prevents current flow into the second body when the at least one switch NMOSFET is enabled.

8. The RF switch of claim 2, wherein the at least one shunt NMOSFET includes a gate, and wherein an electrical device is coupled between the gate and the first ACS such that the electrical device prevents current flow into the first body when the at least one shunt NMOSFET is enabled.

9. The RF switch of claim 2, wherein the couplings comprise at least one of the following coupling types: capacitive; inductive; electro-magnetic; and resistive coupling, including a “short” resistive coupling wherein the resistive value of the coupling comprises approximately zero ohms.

10. The RF switch of claim 2, wherein the first and second negative bias voltages are substantially more negative than the lowest voltage value of ground, a threshold voltage (Vth), a source voltage (Vs), and a drain voltage (Vd).

11. The RF switch of claim 2, wherein the first and second negative bias voltages are at least one volt more negative than the lowest voltage value of ground, a threshold voltage (Vth), a source voltage (Vs), and a drain voltage (Vd).

12. The RF switch of claim 2, wherein each NMOSFET has an associated and respective gate voltage Vg, and wherein the associated gate voltage is at least one volt more negative than the lowest voltage value of ground, a threshold voltage (Vth), a source voltage (Vs), and a drain voltage (Vd).

13. The RF switch of claim 2, further comprising:

a third RF port; and

a second switch transistor grouping coupled with the second and third RF ports wherein the second switch transistor grouping comprises a second plurality of switch NMOSFETs arranged in a stacked configuration, wherein at least one of the second plurality of switch NMOSFETs has a third body and a third accumulated charge sink (ACS) coupled to the third body and configured so that when the at least one switch NMOSFET of the second plurality of switch NMOSFETs is disabled, a third negative bias voltage that is substantially negative with respect to ground applied to the third ACS substantially prevents accumulated charge from accumulating in the third body.

14. The RF switch of claim 13, wherein preventing accumulated charge from accumulating in the third body of the at least one switch NMOSFET improves linearity of the signal passed through to the second RF port.

15. The RF switch of claim 1, wherein the shunt transistor grouping comprises at least three NMOSFETs.

16. The RF switch of claim 1, wherein the shunt transistor grouping comprises at least six NMOSFETs.

17. The RF switch of claim 1, wherein the shunt transistor grouping comprises at least nine NMOSFETs.

18. The RF switch of claim 2, wherein the first switch transistor grouping comprises at least three NMOSFETs.

19. The RF switch of claim 2, wherein the first switch transistor grouping comprises at least six NMOSFETs.

20. The RF switch of claim 2, wherein the first switch transistor grouping comprises at least nine NMOSFETs.

21. The RF switch of claim 2, wherein the first switch transistor grouping comprises at least twelve NMOSFETs.

22. The RF switch of claim 2, wherein the first switch transistor grouping comprises at least fifteen NMOSFETs.

23. The RF switch of claim 13, wherein the second switch transistor grouping comprises at least three NMOSFETs.

24. The RF switch of claim 13, wherein the second switch transistor grouping comprises at least six NMOSFETs.

25. The RF switch of claim 13, wherein the second switch transistor grouping comprises at least nine NMOSFETs.

26. The RF switch of claim 13, wherein the second transistor grouping comprises at least twelve NMOSFETs.

27. The RF switch of claim 13, wherein the second switch transistor grouping comprises at least fifteen NMOSFETs.

28. The RF switch of claim 2, wherein the at least one shunt NMOSFET further comprises an electrical contact region proximate to the first ACS, and wherein the electrical contact region facilitates coupling to the first body.

29. The RF switch of claim 2, wherein the at least one switch NMOSFET further comprises an electrical contact region proximate to the second ACS, and wherein the electrical contact region facilitates coupling to the second body.

30. The RF switch of claim 13, wherein the at least one second switch NMOSFET further comprises an electrical contact region proximate to the third ACS, and wherein the electrical contact region facilitates coupling to the third body.

31. The RF switch according to claim 1, wherein the first switch control signal enabling the first plurality of switch NMOSFETs has a voltage of approximately +2.5 Volts, and wherein the switch and shunt control signals shunt control signal enabling the plurality of shunt NMOSFETs have has a voltage of approximately +2.5 Volts.

32. The RF switch according to claim 1, wherein the first switch control signal disabling the first plurality of switch NMOSFETs has a voltage of approximately −2.5 Volts, and wherein the switch and shunt control signals shunt control signal disabling the plurality of shunt NMOSFETs have has a voltage of approximately −2.5 Volts.

33. The RF switch according to claim 1, wherein the NMOSFETs comprise partially depleted NMOSFETs.

34. The RF switch according to claim 1, wherein the NMOSFETs comprise fully depleted NMOSFETs.

35. The RF switch of claim 1, wherein the silicon layer of the silicon-on-insulator substrate has a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

36. The RF switch of claim 2, wherein the second body has a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

37. The RF switch of claim 13, wherein the third body has a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

38. The RF switch of claim 10, wherein the silicon layer of the silicon-on-insulator substrate has a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

39. The RF switch of claim 11, wherein the silicon layer of the silicon-on-insulator substrate has a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

40. The RF switch of claim 12, wherein the silicon layer of the silicon-on-insulator substrate has a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

41. The RF switch of claim 1, wherein the first switch transistor grouping and the shunt transistor grouping are fabricated in a silicon layer on an insulating layer, and wherein the switch NMOSFETs and shunt NMOSFETs have associated and corresponding sources and drains, and wherein the sources and drains of the switch and shunt NMOSFETs extend substantially proximate the insulating layer.

42. The RF switch of claim 2, wherein the first switch transistor grouping and the shunt transistor grouping are fabricated in a silicon layer on an insulating layer, and wherein the switch NMOSFETs and shunt NMOSFETs have associated and corresponding sources and drains, and wherein the sources and drains of the switch and shunt NMOSFETs extend substantially proximate the insulating layer.

43. The RF switch of claim 13, wherein the first switch transistor grouping and the shunt transistor grouping are fabricated in a silicon layer on an insulating layer, and wherein the switch NMOSFETs and shunt NMOSFETs have associated and corresponding sources and drains, and wherein the sources and drains of the switch and shunt NMOSFETs extend substantially proximate the insulating layer.

44. The RF switch of claim 10, wherein the first switch transistor grouping and the shunt transistor grouping are fabricated in a silicon layer on an insulating layer, and wherein the switch NMOSFETs and shunt NMOSFETs have associated and corresponding sources and drains, and wherein the sources and drains of the switch and shunt NMOSFETs extend substantially proximate the insulating layer.

45. The RF switch of claim 11, wherein the first switch transistor grouping and the shunt transistor grouping are fabricated in a silicon layer on an insulating layer, and wherein the switch NMOSFETs and shunt NMOSFETs have associated and corresponding sources and drains, and wherein the sources and drains of the switch and shunt NMOSFETs extend substantially proximate the insulating layer.

46. The RF switch of claim 12, wherein the first switch transistor grouping and the shunt transistor grouping are fabricated in a silicon layer on an insulating layer, and wherein the switch NMOSFETs and shunt NMOSFETs have associated and corresponding sources and drains, and wherein the sources and drains of the switch and shunt NMOSFETs extend substantially proximate the insulating layer.

47. A circuit routing RF signals, comprising:

a first RF port;

a second RF port;

a third RF port;

a plurality of stacked shunt NMOSFETs coupled with the first RF port and ground, and controlled by a shunt control signal, wherein at least one of the plurality of stacked shunt NMOSFETs comprises a first body and a first accumulated charge sink (ACS) coupled with the first body and configured so that when the at least one shunt NMOSFET is disabled, a first negative bias voltage that is substantially negative with respect to ground applied to the first ACS substantially prevents accumulated charge from accumulating in the first body of the at least one shunt NMOSFET;

a first plurality of first stacked switch NMOSFETs coupled with the first and second RF ports, and controlled by a first switch control signal, wherein at least one of the first plurality of first switch NMOSFETs has a second body and a second accumulated charge sink (ACS) coupled with the second body and configured so that when the at least one first switch NMOSFET is disabled, a second negative bias voltage that is substantially negative with respect to ground applied to the second ACS substantially prevents accumulated charge from accumulating in the second body of the at least one first switch NMOSFET;

a second plurality of second stacked switch NMOSFETs coupled with the second and third RF ports, wherein at least one of the second plurality of second switch NMOSFETs has a third body and a third accumulated charge sink (ACS) coupled to the third body and configured so that when the at least one second switch NMOSFET is disabled, a third negative bias voltage that is substantially negative with respect to ground applied to the third ACS substantially prevents accumulated charge from accumulating in the third body; and

a silicon-on-insulator substrate, wherein the shunt, first switch and second switch NMOSFETs are fabricated in a silicon layer of the silicon-on-insulator substrate.

48. The circuit of claim 47, wherein when the first switch NMOSFETs are enabled by the first switch control signal, the shunt NMOSFETs are disabled by the shunt control signal and the at least one second switch NMOSFET is disabled, a signal on the first RF port is passed through to the second RF port, and wherein preventing accumulated charge from accumulating in the third body of the at least one second switch NMOSFET improves linearity of the signal passed through to the second port.

49. The circuit of claim 47, wherein the silicon layer of the silicon-on-insulator substrate has a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

50. The circuit of claim 48, wherein the silicon layer of the silicon-on-insulator substrate has a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

51. An RF switch, comprising:

a first RF port;

a second RF port;

a first switch transistor grouping coupled to the first RF port and to the second RF port and comprising a first plurality of switch NMOSFETs arranged in a stacked configuration;

a first shunt transistor grouping coupled to the first RF port and to ground and comprising a first plurality of shunt NMOSFETs arranged in a stacked configuration,

wherein at least one of the first plurality of shunt NMOSFETs comprises a first gate, a first source, a first drain, a first body, and a first accumulated charge sink (ACS) coupled to the first body,

wherein a first ACS bias voltage is applied to the first ACS,

wherein the first ACS is in electrical communication with the first body and is configured so that when the at least one shunt NMOSFET of the first plurality of shunt NMOSFETs is operated in an off-state (non-conducting state), the first ACS bias voltage is substantially negative with respect to ground to substantially prevent accumulated charge from accumulating in the first body of the at least one shunt NMOSFET; and

wherein the first switch transistor grouping and the first shunt transistor grouping are fabricated on a silicon-on-insulator substrate.

52. The RF switch of claim 51, wherein preventing accumulated charge from accumulating in the first body of the at least one shunt NMOSFET improves linearity of the at least one shunt NMOSFET.

53. The RF switch of claim 52, wherein the first ACS is positioned proximate a first distal end of the first body.

54. The RF switch of claim 52, wherein the at least one shunt NMOSFET of the first plurality of shunt NMOSFETs further comprises a first electrical contact region, wherein the first electrical contact region and the first ACS are coterminous.

55. The RF switch of claim 54, wherein the first electrical contact region is positioned proximate to and in electrical contact with the first ACS.

56. The RF switch of claim 52, wherein the first ACS is electrically coupled to the gate.

57. The RF switch of claim 52, wherein the first ACS bias voltage applied to the first ACS is negative with respect to ground and has a magnitude of approximately 2.5V in an off-state (non-conducting state).

58. The RF switch of claim 57, wherein the first ACS bias voltage applied to the first ACS has a voltage level in an on-state (conducting state) that is different from a voltage level in an off-state (non-conducting state).

59. The RF switch of claim 51, further comprising:

a third RF port;

a second switch transistor grouping coupled to the third RF port and to the second RF port and comprising a second plurality of switch NMOSFETs arranged in a stacked configuration,

wherein at least one of the second plurality of switch NMOSFETs comprises a second gate, a second source, a second drain, a second body, and a second accumulated charge sink (ACS) coupled to the second body,

wherein a second ACS bias voltage is applied to the second ACS,

wherein the second ACS is in electrical communication with the second body and is configured so that when the at least one switch NMOSFET of the second plurality of switch NMOSFETs is operated in an off-state (non-conducting state), the second ACS bias voltage is substantially negative with respect to ground to substantially prevent accumulated charge from accumulating in the second body of the at least one switch NMOSFET of the second plurality of switch NMOSFETs; and

wherein the second switch transistor grouping is fabricated on the silicon-on-insulator substrate.

60. The RF switch of claim 59, wherein preventing accumulated charge from accumulating in the second body of the at least one switch NMOSFET of the second plurality of switch NMOSFETs improves linearity of the a signal passed through to the second RF port.

61. The RF switch of claim 60, wherein the first ACS is positioned proximate a first distal end of the first body and the second ACS is positioned proximate a second distal end of the second body.

62. The RF switch of claim 61, wherein the a first electrical contact region is positioned proximate to and in electrical contact with the first ACS, and wherein the a second electrical contact region is positioned proximate to and in electrical contact with the second ACS.

63. The RF switch of claim 60, wherein the first ACS and second ACS are is electrically coupled to the first gate, and wherein the second ACS is electrically coupled to the second gate.

64. The RF switch of claim 60, wherein the first ACS bias voltage applied to the first ACS and the second ACS bias voltage applied to the second ACS are negative with respect to ground and have a magnitude of approximately 2.5V in an off-state (non-conducting state).

65. The RF switch of claim 64, wherein the first ACS bias voltage applied to the first ACS and the second ACS bias voltage applied to the second ACS have a voltage level in an on-state (conducting state) that is different from a voltage level in an off-state (non-conducting state).

66. The RF switch of claim 60, wherein the first ACS bias voltage and the second ACS bias voltage have the same magnitude.

67. The RF switch of claim 51, wherein the first switch transistor grouping and the first shunt transistor grouping are fabricated in a silicon layer having a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

68. The RF switch of claim 51, wherein the first switch transistor grouping and the first shunt transistor grouping are fabricated in a silicon layer on the silicon-on-insulator substrate, and wherein the switch NMOSFETs and shunt NMOSFETs have associated and corresponding sources and drains, and wherein the sources and drains of the switch and shunt NMOSFETs extend substantially proximate the substrate.

69. The RF switch of claim 51, wherein the first body has a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

70. The RF switch of claim 59, wherein the second body has a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

71. An RF switch, comprising:

a first RF port;

a second RF port;

wherein a first ACS bias voltage is applied to the first ACS,

wherein the first ACS is in electrical communication with the first body and is configured so that when the at least one shunt NMOSFET of the first plurality of shunt NMOSFETs is operated in an off-state (non-conducting state), the first ACS bias voltage is substantially more negative than the lesser of a first source bias voltage applied to the first source and a first drain bias voltage applied to the first drain to substantially prevent accumulated charge from accumulating in the first body of the at least one shunt NMOSFET; and

72. The RF switch of claim 71, wherein preventing accumulated charge from accumulating in the first body of the at least one shunt NMOSFET improves linearity of the at least one shunt NMOSFET.

73. The RF switch of claim 72, wherein the first ACS is positioned proximate a first distal end of the first body.

74. The RF switch of claim 72, wherein the first ACS is electrically coupled to the gate.

75. The RF switch of claim 72, wherein the first ACS bias voltage applied to the first ACS is negative with respect to ground and has a magnitude of approximately 2.5V in an off-state (non-conducting state).

76. The RF switch of claim 75, wherein the first ACS bias voltage applied to the first ACS has a voltage level in an on-state (conducting state) that is different from a voltage level in an off-state (non-conducting state).

77. The RF switch of claim 71, further comprising:

a third RF port;

wherein a second ACS bias voltage is applied to the second ACS,

wherein the second ACS is in electrical communication with the second body and is configured so that when the at least one switch NMOSFET of the second plurality of switch NMOSFETs is operated in an off-state (non-conducting state), the second ACS bias voltage is substantially more negative than the lesser of a second source bias voltage applied to the second source and a second drain bias voltage applied to the second drain to substantially prevent accumulated charge from accumulating in the second body of the at least one switch NMOSFET of the second plurality of switch NMOSFETs; and

78. The RF switch of claim 77, wherein preventing accumulated charge from accumulating in the second body of the at least one switch NMOSFET of the second plurality of switch NMOSFETs improves linearity of the a signal passed through to the second RF port.

79. The RF switch of claim 78, wherein the first ACS is positioned proximate a first distal end of the first body and the second ACS is positioned proximate a second distal end of the second body.

80. The RF switch of claim 78, wherein the first ACS and second ACS are is electrically coupled to the first gate, and wherein the second ACS is electrically coupled to the second gate.

81. The RF switch of claim 78, wherein the first ACS bias voltage applied to the first ACS and the second ACS bias voltage applied to the second ACS are negative with respect to ground and have a magnitude of approximately 2.5V in an off-state (non-conducting state).

82. The RF switch of claim 81, wherein the first ACS bias voltage applied to the first ACS and the second ACS bias voltage applied to the second ACS have a voltage level in an on-state (conducting state) that is different from a voltage level in an off-state (non-conducting state).

83. The RF switch of claim 78, wherein the first ACS bias voltage and the second ACS bias voltage have the same magnitude.

84. The RF switch of claim 71, wherein the first switch transistor grouping and the first shunt transistor grouping are fabricated in a silicon layer having a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

85. The RF switch of claim 71, wherein the first body has a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

86. The RF switch of claim 71, wherein the first switch transistor grouping and the first shunt transistor grouping are fabricated in a silicon layer on the substrate, and wherein the first source and first drain extend substantially proximate the substrate.

87. The RF switch of claim 78, wherein the first switch transistor grouping and the first shunt transistor grouping are fabricated in a silicon layer having a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

88. An RF switch, comprising:

a first RF port;

a second RF port;

wherein at least one of the first plurality of shunt NMOSFETs comprises a first gate, a first source, a first drain, a first body, a first accumulated charge sink (ACS) coupled to the first body, wherein the first ACS is positioned proximate a first distal end of the first body and is in electrical communication with the first body, and a second accumulated charge sink (ACS) coupled to the first body, wherein the second ACS is positioned proximate a second distal end of the first body and is in electrical communication with the first body,

wherein a first ACS bias voltage is applied to the first ACS and second ACS,

wherein, when the at least one shunt NMOSFET of the first plurality of shunt NMOSFETs is configured to operate in an off-state (non-conducting state), the first ACS bias voltage is substantially negative with respect to ground to substantially prevent accumulated charge from accumulating in the first body of the at least one shunt NMOSFET; and

89. The RF switch of claim 88, wherein preventing accumulated charge from accumulating in the first body of the at least one shunt NMOSFET of the first plurality of shunt NMOSFETs improves linearity of the at least one shunt NMOSFET.

90. The RF switch of claim 89, further comprising:

a first electrical contact region positioned proximate to and in electrical communication with the first ACS, wherein the first electrical contact region facilitates electrical coupling to the first ACS; and

a second electrical contact region positioned proximate to and in electrical communication with the second ACS, wherein the second electrical contact region facilitates electrical coupling to the second ACS.

91. The RF switch of claim 90, wherein the first ACS and second ACS are electrically coupled to the gate.

92. The RF switch of claim 89, wherein the first ACS bias voltage applied to the first ACS and the second ACS are negative with respect to ground and have a magnitude of approximately 2.5V in an off-state (non-conducting state).

93. The RF switch of claim 92, wherein the first ACS bias voltage applied to the first ACS and the second ACS has a voltage level in an on-state (conducting state) that is different from a voltage level in an off-state (non-conducting state).

94. The RF switch of claim 88, wherein the first switch transistor grouping and the first shunt transistor grouping are fabricated in a silicon layer having a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

95. The RF switch of claim 88, wherein the first body has a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

96. The RF switch of claim 88, wherein the first switch transistor grouping and the first shunt transistor grouping are fabricated in a silicon layer on the substrate, and wherein the first source and first drain extend substantially proximate the substrate.

97. An RF switch, comprising:

a first RF port;

a second RF port;

wherein a first ACS bias voltage is applied to the first ACS and second ACS,

wherein, when the at least one shunt NMOSFET of the first plurality of shunt NMOSFETs is configured to operate in an off-state (non-conducting state), the first ACS bias voltage is substantially more negative than the lesser of a first source bias voltage applied to the first source and a first drain bias voltage applied to the first drain to substantially prevent accumulated charge from accumulating in the first body of the at least one shunt NMOSFET; and

98. The RF switch of claim 97, wherein preventing accumulated charge from accumulating in the first body of the at least one shunt NMOSFET of the first plurality of shunt NMOSFETs improves linearity of the at least one shunt NMOSFET.

99. The RF switch of claim 98, further comprising:

100. The RF switch of claim 99, wherein the first ACS and second ACS are electrically coupled to the gate.

101. The RF switch of claim 98, wherein the first ACS bias voltage applied to the first ACS and the second ACS has a voltage level in an on-state (conducting state) that is different from a voltage level in an off-state (non-conducting state).

102. The RF switch of claim 97, wherein the first switch transistor grouping and the first shunt transistor grouping are fabricated in a silicon layer having a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

103. The RF switch of claim 98, wherein the first body has a thickness between approximately 100 angstroms to approximately 2,000 angstroms.

104. The RF switch of claim 97, wherein the first switch transistor grouping and the first shunt transistor grouping are fabricated in a silicon layer on the substrate, and wherein the first source and first drain extend substantially proximate the substrate.