US20100103578A1 - Electrical stress protection apparatus and method of manufacture - Google Patents

Abstract

In various embodiments, circuits and semiconductor devices and structures and methods to manufacture these structures and devices are disclosed. In one embodiment, a bidirectional polarity, voltage transient protection device is disclosed. The voltage transient protection device may include a bipolar PNP transistor having a turn-on voltage of V_BE1, a bipolar NPN transistor having a turn-on voltage of V_BE2, and a field effect transistor (FET) having a threshold voltage of V_TH, wherein a turn-on voltage V_TOof the voltage transient protection device is approximately equal to the sum of V_BE1, V_BE2, and V_TH, that is, V_TO≅V_BE1+V_BE2+V_TH. Other embodiments are described and claimed.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS
• The present application is a divisional of U.S. application Ser. No. 11/467,452 filed Aug. 25, 2006. Said application Ser. No. 11/467,452 is hereby incorporated herein by reference in its entirety.
TECHNICAL FIELD
• Embodiments disclosed in the present disclosure relate generally to electrical and semiconductor technology, and more specifically to circuits and semiconductor devices for reducing negative effects associated with electrical stresses and methods of their manufacture
BACKGROUND
• Active electronic components and integrated circuits are increasingly significant in modern controllers, communications equipment and related or compatible systems. Electronic component manufacturers are constantly striving to increase the performance of their products, while decreasing their cost of manufacture. Economic concerns and market forces driven by larger system and computation desires result in the desire for increasing circuit complexity and breadth of functionality. These concerns and forces, including size and power efficiency considerations, may place constraints on the elements and functions that are combined in realization of such circuitry.
• General trends towards progressively smaller devices and reduced power consumption per circuit element may result in increased susceptibility of these devices to catastrophic failure. Electrical stresses may originate from a variety of different phenomena, including electrostatic discharge (ESD) from environmental sources, voltage stresses originating from switching and other electronic functions in such circuitry, or circuitry coupled thereto, and may be exacerbated by ringing in electrical signal and power distribution and coupling circuitry, ground or other power-supply conductor potential disturbances, or by failure or malfunction of portions of circuitry coupled to an affected component. Susceptibility of electrical components to effects of electrical stress may increase as size of individual elements decreases, in part because progressively lower power supply voltages are consistent with these trends, and, as a result, components in these circuits are increasingly voltage-sensitive. Accordingly, these concerns collectively present competing challenges, particularly in view of performance targets for robustness and reliability of resultant electronic circuits.
• Different surge or transient suppressor devices and designs have been developed, responsive to long-felt needs within the industry. Some approaches require relatively large areas for formation or may require additional processing considerations or fabrication elements such as elements or processing considerations for masking operations. Other approaches may present parasitic electrical effects that in turn may affect circuit performance adversely or have performance characteristics susceptible to degradation or catastrophic failure in their intended application. In some approaches, provision of capacity for carrying sufficient electrical current may also result in unwieldy footprint requirements or reduced switching speed or both.
BRIEF DESCRIPTION OF THE DRAWINGS
• FIG. 1 is a block diagram of an electrical transient suppression device;
• FIG. 2 is a graph displaying a piecewise-linear approximation to current (ordinate) versus voltage (abscissa) for the transient suppression element ofFIG. 1;
• FIG. 3 is a graph displaying a piecewise-linear approximation of voltage (ordinate) versus time (abscissa) relevant to the environment ofFIG. 1;
• FIG. 4 is a graph displaying a piecewise-linear approximation of current (ordinate) versus time (abscissa) relevant to the environment ofFIG. 1;
• FIG. 5 is an exemplary schematic diagram of an embodiment of the transient suppression element ofFIG. 1;
• FIG. 6 is a schematic diagram of an embodiment of the transient suppression device ofFIG. 1;
• FIG. 7 is a plan view of an embodiment of a circuit layout capable of use with the transient suppression device ofFIG. 1;
• FIG. 8 is a cross section view taken along section lines VIII-VIII ofFIG. 7, of a realization of the embodiment depicted schematically inFIG. 6;
• FIG. 9 is an enlarged side view of a portion of the view ofFIG. 8;
• FIG. 10 is a composite physical and schematic diagram of a portion of the structure inFIG. 8;
• FIG. 11 is a cross section view of a portion of a structure at one stage during manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 12 is a cross section view of the structure shown inFIG. 11 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 13 is a cross section view of the structure shown inFIG. 12 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 14 is a cross section view of a portion of a structure at one stage during manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 15 is a cross section view of the structure shown inFIG. 14 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 16 is a cross section view of the structure shown inFIG. 15 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 17 is a cross section view of the structure shown inFIG. 16 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 18 is a cross section view of the structure shown inFIG. 17 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 19 is a cross section view of the structure shown inFIG. 18 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 20 is a cross section view of the structure shown inFIG. 19 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 21 is a cross section view of the structure shown inFIG. 20 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 22 is a cross section view of the structure shown inFIG. 21 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 23 is a cross section view of the structure shown inFIG. 22 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 24 is a cross section view of the structure shown inFIG. 23 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 25 is a cross section view of the structure shown inFIG. 24 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 26 is a cross section view of the structure shown inFIG. 25 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 27 is a cross section view of the structure shown inFIG. 26 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 28 is a cross section view of the structure shown inFIG. 27 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 29 is a cross section view of a structure in accordance with an embodiment of the present disclosure;
• FIG. 30 is a cross section view of a portion of a structure at one stage during manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 31 is a cross section view of the structure shown inFIG. 30 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 32 is a cross section view of the structure shown inFIG. 31 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 33 is a cross section view of the structure shown inFIG. 32 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 34 is a cross section view of the structure shown inFIG. 33 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure;
• FIG. 35 is a cross section view of a stand-alone component embodiment incorporating at least one transient suppression device of the present disclosure;
• FIG. 36 is a simplified plan view of an integrated circuit embodiment incorporating one or more of the transient suppression devices of the present disclosure; and
• FIG. 37 is a simplified schematic diagram of a RF amplifier embodiment including at least one transient suppression device of the present disclosure.
• For simplicity of illustration and ease of understanding, elements in the various figures are not necessarily drawn to scale, unless explicitly so stated. In some instances, well-known methods, procedures, components and circuits have not been described in detail so as not to obscure the present disclosure. The following detailed description is merely exemplary in nature and is not intended to limit the disclosure of this document and uses of the disclosed embodiments. Furthermore, there is no intention to be bound by any expressed or implied theory presented in the preceding text, including the title, technical field, background, or the following abstract.
DETAILED DESCRIPTION
• In the following description and claims, the terms “comprise” and “include,” along with their derivatives, may be used and are intended as synonyms for each other. In addition, in the following description and claims, the terms “coupled” and “connected,” along with their derivatives, may be used. It should be understood that these terms are not necessarily intended as synonyms for each other. Rather, in particular embodiments, “connected” may be used to indicate that two or more elements are in direct physical or electrical contact with each other. “Coupled” may mean that two or more elements are in direct physical or electrical contact. However, “coupled” may also mean that two or more elements are not in direct contact with each other, but yet still co-operate or interact with each other. For example, “coupled” may mean that two or more elements do not contact each other but are joined together via another element or intermediate elements.
• Transistors may be referred to generally as active elements and resistors, inductors, and capacitors may be referred to generally as passive elements. As is generally understood, a bipolar transistor includes a collector region, a base region, and an emitter region and a field effect transistor (FET) includes a gate, a drain region, a source region, and a channel region. The drain region, the source region, the channel region, or the gate of a FET may each be referred to as a portion, a part, a component, or an element of the FET, and similarly, the collector region, the base region, and the emitter region of a bipolar transistor may each be referred to as a portion, a part, a component, or an element of the bipolar transistor.
• Generally, transistors such as bipolar transistors and field effect transistors (FETs) discussed herein are understood to provide a conduction path between first and second conduction electrodes when a control signal is applied to a control electrode. For example, in a FET a channel region formed between the drain and source electrodes provides the conduction path which is controlled in accordance with the magnitude of the control signal. In addition, the conduction path of a MOS transistor can be enabled by applying a voltage in excess of the drain-to-source breakdown voltage of the FET. The gate electrode of a FET may be referred to as a control electrode and the drain and source electrodes of a FET may be referred to as current carrying electrodes or conduction electrodes. Likewise, the base of a bipolar transistor may be referred to as the control electrode and the collector and emitter electrodes of the bipolar transistor may be referred to as conduction electrodes or current carrying electrodes. In addition, the drain and source electrodes of a FET may be referred to as power electrodes and the collector and emitter electrodes of a bipolar transistor may also be referred to as power electrodes.
• The following disclosure describes improved apparatus, techniques and processes for design and fabrication of circuit elements for reducing destructive aspects of electrical stress effects. The disclosure focuses on such in the context of such devices formed using monolithic single crystal silicon substrates, however, it will be appreciated that the concepts contained herein have application for other types of circuit elements and are useful in the context of other types of conductive or dielectric substrate materials such as, for example, gallium arsenide (GaAs).
• Architectures and methodologies relevant to microelectronic apparatus are disclosed. In a first aspect, the present disclosure contemplates a voltage clipping mechanism. The voltage clipping mechanism includes a bipolar multilayer switching region and a metal-oxide-semiconductor field effect transistor (MOSFET) region. The MOSFET region has a predetermined threshold voltage V_TH. A turn-ON voltage V_ONof the voltage clipping mechanism is given by: V_ON=V_TH+2V_BE, where V_BErepresents turn-ON voltage for p-n junctions in the multilayer switching region.
• In another aspect, the disclosed concepts include a voltage stress protection structure. The device includes a semiconductive substrate doped to provide a relatively high conductivity and a clamp circuit formed over the substrate. The clamp circuit includes first and second electrodes and in some embodiments, is capable of being configured to provide a clamp voltage in a range of about six or about seven volts, or less, with respect to a voltage difference between the first and second electrodes. In some embodiments, the clamp circuit includes clamping characteristics that are symmetric about a predetermined or an arbitrarily-chosen voltage as is described below.
• The following section, includingFIGS. 1 through 4, addresses an overview of an architecture for utilization of improved electrical stress reduction structures or voltage clipping mechanisms, including such structures when integrated in conjunction with electrical and/or semiconductive devices, in accordance with the teachings of the disclosure. The disclosed apparatuses are designed to fulfill multiple areas of functionality and to satisfy a plurality of performance targets simultaneously, while also conforming with many other criteria (e.g., manufacturability). In various embodiments, the disclosed apparatuses are non-linear elements intended to ensure that undesirable electrical stress such as, for example, undesirable voltages (such as voltages induced by electrostatic discharge, electromagnetic fields, or other voltage, current or charge transients) are not impressed on vulnerable elements. In some embodiments discussed below, a non-linear element is disclosed that provides a relatively high-speed conduction path for such charge or electrical stress and yet does not provide undesirable RF characteristics, such as relatively high shunt impedance. The concepts disclosed herein are capable of voltage clamping characteristics over a broad range of voltages, and in some embodiments, also provide effective voltage stress protection in a range of six or seven volts or less, together with appropriately low OFF state parasitic effects and clamping characteristics that are symmetric about an arbitrarily-chosen voltage. In many situations, some of the apparatuses disclosed herein may pass relatively little quiescent current until a voltage threshold is reached, and then may pass substantially greater currents, as will be explained below in more detail.
• FIG. 1 is a block diagram100 of an electricaltransient suppression device105 coupled to an environment via afirst terminal110 and asecond terminal115 and/or to electronic circuitry (not illustrated inFIG. 1). In various embodiments,terminals110 and115 may be referred to as conductors anddevice105 may be referred to as an electrical stress protection apparatus or element, electrical stress reduction structure, electrical stress relief structure, voltage clipping mechanism, voltage stress protection device, transient suppression device, surge or transient suppressor device, voltage transient surge protection device, voltage transient protection device, clamp circuit, clipper circuit, limiter circuit, or a non-linear element. In addition, in some embodiments,device105 may be referred to as an ESD protection device, structure, circuit, or component and may be a relatively fast, bidirectional ESD protection device in some embodiments.
• Device105 has twoterminals110 and115 respectively coupled to a firstelectrical conductor120 and a secondelectrical conductor125. The first andsecond conductors120 and125, in turn, may form at least one electrical port, such asport130, and may also form a second port135 (each represented in part by a dashed vertical line), labeled as presenting impedances Z_INand Z_OUT, respectively.
• Although the scope of the present disclosure is not limited in this respect,electrical conductors120 and125 may be power distribution or other conductors formed in monolithic circuitry, interconnections employed in hybrid circuitry multi-chip technologies, where multiple devices are connected together to form a module, or may be external signal conduction paths or other electrical conductors. For example,device105 may be coupled between two traces formed in a circuit, where the traces comprise a portion of a path for coupling a signal from one portion of circuitry to another or be associated with input/output structures. These signals may be switched signals, as in clock distribution schemes, or may be radio frequency (RF) signal conduction paths or may serve other purposes. In such instances, both the input impedance Z_INand the output impedance Z_OUTmay be needed for accurate RF modeling of circuit performance, anddevice105 may be modeled as a two-port component, as illustrated inFIG. 1.
• Alternatively,device105 may be coupled in shunt with another circuit element, for example to protect an amplifier input stage, or in shunt with an input/output (I/O) port such as a bond pad or conductive bump interfacing an electrical component or integrated circuit with other elements or may be coupled in shunt with internal but vulnerable points within a device or component. In such cases, one or both of theports130,135 would represent the electrical connections to thedevice105, and one or both of the port impedances Z_IN, Z_OUTmay be needed for accurate RF modeling of circuit performance.
• When avoltage V140 is developed across theterminals110 and115, a current I150 results through thedevice105. This is described in more detail below with reference toFIGS. 2,3 and4.
• FIG. 2 is agraph200 displaying a piecewise-linear approximation to a current-voltage (IV) curve applicable todevice105 ofFIG. 1. The IV curve representation consists oflinear line segments205,205′,210 and210′, depicted with respect tovoltage240 and current250. It will be appreciated that although the IV curve representation ofFIG. 2 is symmetric about the origin, this curve also represents a family of similar curves having symmetry about a point corresponding to zero current together with an arbitrary or predetermined voltage offset. Thedevice105 is assumed to be characterizable in terms of a turn-ON voltage V_TO. For relatively small values ofvoltage V240, that is, a voltage V having a magnitude of V_TOor less, theline segments205 and205′ correspond to a current I250 throughdevice105 having a magnitude of I₀or less. These segments also correspond to relatively high effective direct current (DC) and RF impedance between theterminals110 and115.
• When thevoltage V140 increases above a voltage of V_TOvolts (or decreases below a voltage of −V_TOvolts), current I250 through thetransient suppression device105 abruptly increases to values much greater than I₀(or much more negative than −I₀). In these conduction regimes, the effective impedance decreases extremely sharply in the vicinity of V_TO, as represented by theline segments210 and210′. The turn-ON voltage V_TOthus provides a sharp discrimination between voltages greater and/or smaller than the turn-ON voltage V_TO.
• Thegraph200 ofFIG. 2 displays an IV curve for atransient suppression device105 having symmetric characteristics. It will be appreciated, however, that other combinations of positive and negative characteristics are possible, and are included within the scope of the present disclosure. Although the scope of the present disclosure is not limited in this respect, in some embodiments, the turn-ON voltage V_TOmay be a voltage of less than about seven volts. For example, in various embodiments, the turn-ON voltage V_TOmay be about five volts, two volts, 1.8, or 1.5 volts. Accordingly,device105 provides a relatively low voltage transient suppression device that may be used in low voltage applications or devices.
• FIG. 3 is agraph300 displaying voltage340 (ordinate) versus time360 (abscissa), andFIG. 4 is agraph400 displaying current450 (ordinate) versus time460 (abscissa), relevant to the description of the operation ofdevice105 ofFIG. 1.Voltage340 ofFIG. 3 is analogous to, and can correspond to,voltages140 ofFIGS. 1 and 240 ofFIG. 2, and likewise, current450 ofFIG. 4 is analogous to, and can correspond to, current150 ofFIG. 1. In addition,time360 ofFIG. 3 is analogous to, and can correspond to, time460 ofFIG. 4.
• FIG. 3 combines representations of several different conditions for the purpose of ready comparison. One condition corresponds to a situation where thetransient suppression device105 is combined with several other elements, while a second condition corresponds to an “open circuit” electrical event and hypothetical voltages for a case where thetransient suppression device105 is absent, and it is assumed that the other elements do not significantly alter the open-circuit voltage.
• FIG. 3 illustrates an example, whereinvoltage340 is zero until time T₁. For the first condition, after the time T₁, when an electrical transient is assumed to initiate, thevoltage340 initially increases (solid trace362), at least until a desired turn-ON value V₁is reached at time T₂. Under open-circuit conditions, thevoltage340 also increases and is also represented by theline segment362 at values below the turn-ON value V₁. Voltage V₁ofFIG. 3 is analogous to, and can correspond to, voltage V_TOofFIG. 2. Similarly, Voltage −V₁ofFIG. 3 is analogous to, and can correspond to, voltage −V_TOofFIG. 2. Thevoltage340 is such as might occur from electrostatic discharge, switching transients, power distribution circuitry effects, or a number of causes.
• Referring toFIGS. 1-3, however, when the voltage340 (140,240) across thetransient suppression element105 reaches the value V₁and attempts to increase past the turn-ON voltage V₁(V_TO) of thetransient suppression device105, thetransient suppression device105 abruptly begins to switch to a relatively lower-resistance state, and thevoltage340 is effectively clamped at the value of V₁, as represented by the solidhorizontal trace366, for example, and current through thetransient suppression device105 begins to abruptly increase, as shown bysolid trace470 inFIG. 4. In other words, when thetransient suppression device105 is included, the voltage excursion that would otherwise occur, represented by dashedtraces364,364′,364″, is precluded, with the result that the voltage at theterminals110 and115 may be as represented by thesolid traces362,366,362′,366′,362″ and the corresponding currents represented by thesolid traces470,470′ and470″ with timing relationships as illustrated inFIGS. 3 and 4.
• In many applications, including some applications in high frequency amplifiers, a direct-current steady-state bias voltage, such as is represented by V_BinFIG. 3, may be impressed across theterminals110 and115. In RF amplifier input applications, a DC input bias voltage V_Bof one to several volts may be desired. In some other applications, such as where atransient suppression device105 is employed to dissipate power distribution circuitry transients, a power signal (such as V_DD) may be impressed on one of theterminals110, while a ground connection or signal return path may be supplied via theother terminal115.
• It may be desirable to provide an automatic resetting capability for thetransient suppression device105, thus limiting disruption of normal circuit operation to an interval determined primarily by the duration T_D(FIG. 3) of the transient electrical disturbance. In the scenario depicted inFIG. 3, a superposition of the bias voltage V_Band a signal peak-to-peak voltage V_PPas applied to theterminals110 and115 of thetransient suppression device105 is represented by the bracketed region that is vertically bisected by the horizontal trace associated with the bias voltage V_B, as shown inFIG. 3.
• In order for a high frequency signal such as an input signal to an amplifier to provide resetting capability, some factors may be significant, for example: (i) control of the threshold or thresholds at which transient suppression operation begins and ends, and (ii) an absence of “snap-back” in the operating characteristics of thetransient suppression device105. In a context relevant to many applications, stress responsive elements such asdevice105 providing resetting capability unresponsive to voltages within a first range of voltages (such as exemplified via V_PPinFIG. 3) but having reset capabilities (manifesting return to the OFF mode) associated with voltages adjacent to that first voltage range, and also coinciding with a range of voltages falling within a range associated with substantially less risk of electrically-related stress yet adjacent to the first voltage range and having a narrowly-defined breadth outside that of the first voltage range by a factor such as, for example, ten percent or less, may be desirable. In this scenario, where resetting responsive to signal conditions relevant to design and operation are considerations, but where simplified operation is also desirable, transient suppression apparatus requiring resetting operations to obviate latching phenomena resulting from negative resistance, or snap-back, may be undesirable and may present performance limitations inconsistent with intended purposes. In some embodiments,device105 described herein, is configured or adapted to provide a current-voltage characteristic exhibiting only positive effective direct current resistance at theterminals110 and115 so that there is an absence of “snap-back” in the operating characteristics of thetransient suppression device105.
• Referring again toFIGS. 1-4, snap-back is a phenomenon wherein a first characteristic voltage V₁applied acrossterminals110 and115 initiates conduction of relatively significant current150 in thetransient suppression device105, and a second, relatively lower characteristic voltage V₂is relevant to describing a later portion of conduction of thetransient suppression device105. For example, in some types oftransient suppression devices105, the first characteristic voltage V₁may be approximated as 2V_FB+V_THRES, while the second characteristic voltage V₂may be approximated as V_FB+V_SAT, where V_FBrepresents a voltage drop typical of a forward-biased diode, V_THRESrepresents a threshold voltage of a switching element within thetransient suppression device105 and V_SATrepresents a saturated voltage associated with the switching element, and where V_SATmay be substantially less than V_THRES. Other types of voltages descriptive of thetransient suppression element105 may lead to the same net situation, where the steady-state voltage or second characteristic voltage V₂is substantially less than the initial or first characteristic voltage V₁.
• In a situation where thetransient suppression device105 is coupled in shunt with an input to a high-frequency amplification element, the input signal excursions may be on the order of about one volt, and have frequencies in a range including several tens of gigahertz (GHz) or even higher frequencies. In this scenario, the portion of an RF signal having an amplitude of about one volt which could automatically turn off or reset thetransient suppression device105 to a non-conducting state may have a time duration of picosecond or less. Thus, as an example, when an RF device such as an amplifier requires a DC bias of one and a half volts, the first characteristic voltage V₁may be several volts, for example, about three volts, and the second characteristic voltage V₂may be substantially less, for example, about nine-tenths of a volt or less, once thetransient suppression device105 has been triggered, the input signal alone likely will not automatically reset thetransient suppression device105 to the non-conducting state.
• Alternatively, even when the voltage requirements are met, turn-OFF speed of thetransient suppression device105 may be grossly inadequate for purposes of resetting thetransient suppression device105 to a non-conducting or OFF state. When the turn-OFF switching speed for thetransient suppression device105 requires substantially greater time than a duration of the portion of the RF signal having appropriate voltage excursion characteristics, thetransient suppression device105 very likely cannot be automatically reset to the non-conducting state by the RF signal, even when a bottom edge of the superposed bias voltage V_Bplus signal voltage V_PPincludes excursions below the voltage V₂.
• In such cases, either the ensemble of components must be reset by resetting, for example, removing, the power supply signals or voltages, including the DC bias component, or thetransient suppression device105 must be reset via activation of additional suitable circuitry. Neither of these options is attractive from the perspectives of added device complexity and footprint or of robust and automatic circuit operation.
• In summary, the preceding section describes an environment in which atransient suppression device105 finds utility, and describes properties of thetransient suppression device105 as observed via the terminals to the transient suppression device. The terminal properties described include nonlinear current vs. voltage characteristics and voltage and current versus time performance of at least one type oftransient suppression device105. The description focuses on how the terminal properties interact with the environment, and particularly on how turn-ON and turn-OFF characteristics, as well as I-V characteristics, affect performance and range of useful application fortransient suppression devices105. Electrical circuit embodiments oriented towards realization of these terminal properties are described in the following section.
• InFIGS. 5 and 6, two different circuit embodiments of a transient suppression device105 (FIG. 1) are described.FIGS. 5 and 6 provide schematic diagrams of anembodiment500 and anembodiment600, respectively, of thetransient suppression device105 ofFIG. 1.
• Theembodiment500 depicted inFIG. 5 includes afirst conductor510 and asecond conductor515, analogous to theterminals110 and115 ofFIG. 1, for example. InFIG. 5, elements share the same numbers with primes to illustrate analogous elements incircuit500. For example, theNPN transistors540 and540′ ofembodiment500 share the same or similar characteristics, and are symmetrical in some ways, for example, in terms of their structure, electrical properties (for example, current gain) and in the way these transistors are connected to other elements. Similarly,resistor568,transistor560, andtransistor580 are respectively analogous toresistor568′,transistor560′, andtransistor580′.
• Theembodiment500 includes aNPN transistor540, and aNPN transistor540′, each depicted as having a power electrode oremitter542 and542′, respectively, and each having an associated current gain, β_NPN. The current gain β of a transistor differs from current gain of a transistor amplifier, which is represented as A_I. Current gain β of a transistor in a common-emitter configuration is defined as collector current I_Cdivided by base current I_B, or I_C/I_B.
• TheNPN transistor540 also includes a base orcontrol electrode546 and a collector orsecond power electrode548, while theNPN transistor540′ includes similarly-numbered elements bearing primes, that is,NPN transistor540′ includes acontrol electrode546′ andcollector electrode548′. Theemitter542 of theNPN transistor540 is connected directly to theconductor510, with theNPN transistor540′ being analogously connected to theconductor515, that is, theemitter542′ of theNPN transistor540′ is connected directly to theconductor515.
• Theembodiment500 also includes aPNP transistor560 and aPNP transistor560′, each depicted as havingmultiple emitters562 and561, and562′ and561′, respectively, and each having an associated current gain β_PNP. ThePNP transistor560 includes an emitter orfirst power electrode562, a base orcontrol electrode564 and a collector orsecond power electrode567. Aconductor570, represented as a horizontal central line inFIG. 5, couples thebases564,564′ of the twoPNP transistors560,560′ together. Thecollector567 of thePNP transistor560 is coupled to thebase546 of theNPN transistor540, and both are coupled via aresistor568′ to theconductor510. Similarly, thecollector567′ of thePNP transistor560′ is coupled to the base546′ of theNPN transistor540′, and both are coupled via aresistor568 to theconductor515. Theemitter561 ofPNP transistor560 is connected toconductor515 and theemitter561′ ofPNP transistor560′ is connected toconductor510.
• AMOSFET580 and aMOSFET580′ are also included in theembodiment500 ofFIG. 5, and are depicted inFIG. 5 as being enhancement mode n-channel MOSFETs, that is, for example, having a p-type semiconductor body and operating in an enhancement mode. TheMOSFET580 includes a source orpower electrode582 that is coupled to theemitter542 of theNPN transistor540, and theMOSFET580′ includes a source orpower electrode582′ that is coupled to theemitter542′ of theNPN transistor540′. TheMOSFET580 includes a gate orcontrol electrode584 and a drain orsecond power electrode586 which are coupled together and to theconductor570, and theMOSFET580′ has a gate orcontrol electrode584′ and a drain orsecond power electrode586′ that are coupled together and to theconductor570.
• TheMOSFET580 also includes abody electrode588 that is coupled to thebase546 of theNPN transistor540, to thecollector567 of thePNP transistor560, to theemitter562′ of thePNP transistor560′, and to theresistor568′. TheMOSFET580′ also includes abody electrode588′ that is coupled to the base546′ of theNPN transistor540′, to thecollector567′ of thePNP transistor560′, to theemitter562 of thePNP transistor560, and to theresistor568. EachMOSFET580,580′ has a characteristic threshold voltage V_TH, and a characteristic transconductance g_m, often specified in terms of transconductance per unit of control electrode width. The MOSFET threshold voltages V_THare assumed to be approximately equal in magnitude, and are determinable via factors including doping levels employed during fabrication, as will be subsequently explained in more detail.
• In operation, theembodiment500 has a continuum of stable states, including at least two primary or characterizing operating modes. These primary modes are: (i) an OFF mode, corresponding to trace205 or205′ ofFIG. 2, and (ii) an ON mode, corresponding to trace210 or210′ ofFIG. 2.
• In the OFF mode, there is essentially zero volts applied acrossconductors510 and515, and therefore, thetransistors540,540′,560,560′ and580,580′ are all OFF, that is, relatively little, to no electrical current is conducted via any of the power electrodes. In some embodiments, in the OFF mode, a relatively small amount of current such as quiescent current or leakage current may flow in or throughcircuit500. The OFF mode may be referred to as a non-conduction mode, low-conduction mode, or low current mode of operation.
• For the ON mode, the operation will be described in terms of a first polarity (for example, positive polarity) of voltage140 (FIG. 1),240 (FIG. 2),340 (FIG. 3) that is presented across theconductors510 and515. As will be described in more detail below, in the first polarity example wherein voltage present on theconductor510 increases relative to a voltage present on theconductor515 to provide a voltage increase with respect to time, the transistors having primed reference characters, that is,transistors560′,580′, and540′, will be turned on and active during this mode of operation. In the second polarity example wherein voltage present on theconductor515 increases relative to a voltage present on theconductor510 to provide a voltage increase with respect to time, the transistors having reference numbers without the primed reference characters, that is,transistors560,580, and540, will be turned on and active during this mode of operation. Accordingly, the circuit inembodiment500 has bidirectional and symmetrical characteristics in that it can respond to electrical transients of either polarity applied acrossconductors510 and515.
• After a “turn-on” voltage is reached, that is, after the magnitude, or the absolute value of, the voltage difference of V_TOis applied acrossconductors510 and515,circuit500 will be in the ON mode of operation, wherein a relatively large amount of electrical current may flow in or throughcircuit500, for example, the amount of electrical current flowing incircuit500 is greater than, and in many cases, substantially greater than, leakage or quiescent current amounts. The ON mode may be referred to as a high-conduction mode or high current mode of operation. Accordingly, the amount of electrical current flowing through thedevice105 is less, and in many cases, substantially less, in the OFF mode compared to the amount of electrical current flowing through thedevice105 in the ON mode. In some embodiments, the amount of electrical current flowing through thedevice105 in the OFF mode is approximately zero amperes, or less than about ten microamperes is some cases, and the amount of electrical current flowing through thedevice105 in the ON mode is greater than approximately 200 milliamperes (mA), and in some cases may range from about 200 milliamperes (mA) to several amperes, depending on the size of elements ofstructure800 described herein.
• As an example, as the voltage presented onconductor510 relative toconductor515 rises towards the positive turn-ON voltage V_TOof thetransient suppression device105, this will initiate some relatively small electrical current flow through theemitter561′-base564′ junction of thePNP transistor560′, and throughresistor568′ and then through theemitter562′-base564′ junction of thePNP transistor560′. As will be described in more detail below, in some embodiments, the current gain oftransistor560′ is relatively small compared to, for example, the current gain ofNPN transistor540′, and as a result most of the current throughPNP transistor560′ is base current. As current begins to flow from the base564′, then electrical current will begin to flow through n-channel MOSFET580′, that is through thedrain586′ and thesource582′ ofMOSFET580′. The relatively small current flow throughPNP transistor560′ andMOSFET580′ may be referred to as “leakage current.” As a result, a voltage onconductor570 will begin to rise toward the threshold voltage ofMOSFET580′, and a voltage V_DSdeveloped across theMOSFET580′, from joineddrain586′ andgate584′, to thesource582′, also begins to increase. This drain-source voltage V_DSwill saturate as it reaches the threshold voltage V_THfor theMOSFET580′.
• There is a relatively small amount of current throughMOSFET580′ until the threshold voltage V_THofMOSFET580′ is attained. After the threshold voltage V_THofMOSFET580′ is attained,MOSFET580′ will turn on more significantly to increase the amount of current throughMOSFET580′.
• The current through theemitter562′-base564′ junction of thePNP transistor560′ and theemitter561′-base564′ junction of thePNP transistor560′ gives rise to a current through thecollector567′ (I_C=β_PNP*I_B) and the current fromcollector567′ flows to theresistor568. As a result, a voltage will be generated and applied to the base546′ ofNPN transistor540′ and to thebody electrode588′ ofMOSFET580′, and this voltage will rise, resulting in an increase of the threshold voltage V_THofMOSFET580′ since increasing bias voltage applied to thebody electrode588′ acts to increase the effective threshold voltage V_THof theMOSFET580′. The voltage developed across theresistor568 saturates at a voltage that is determined by the voltage drop across the base546′-emitter542′ junction (V_BE) of theNPN transistor540′, which in some embodiments, is a voltage V_BEranging from about 0.6 volts to about 0.7 volts.
• A relatively small amount of current is flowing throughcircuit500 at this time prior to the time when the voltage difference acrossconductors510 and515 reaches a value approximating a turn-ON voltage V_TOof thetransient suppression device105, which is the summation of the threshold voltage V_THofMOSFET580′ and the turn-ON voltages V_BEofbipolar transistors560′ and 540′. Accordingly, the turn-ON voltage V_TOofcircuit500 ordevice105 may be represented by the equation 2*V_BE+V_TH.
• Theembodiment600 illustrated inFIG. 6 is similar in structure and operation to the embodiment500 (FIG. 5) discussed above. Differences betweenembodiments500 and600 includecircuit600 havingNPN transistors640 and640′, both of which includemultiple emitters642 and644, and642′ and644′, respectively. In addition,circuit600 includesresistors650 and650′.
• Theembodiment600 depicted inFIG. 6 includes afirst conductor610 and asecond conductor615, analogous to theterminals110 and115 ofFIG. 1, for example. InFIG. 6, elements share the same numbers with primes to illustrate analogous elements incircuit600. For example, theNPN transistors640 and640′ ofembodiment600 share the same or similar characteristics, and are symmetrical in some ways, for example, in terms of their structure, electrical properties (for example, current gain) and in the way these transistors are connected to other elements. Similarly,resistor668,resistor650,transistor660, andtransistor680 are respectively analogous toresistor668′,resistor650′,transistor660′, andtransistor680′.
• Theembodiment600 includes aNPN transistor640 and aNPN transistor640′, each depicted as havingmultiple emitters642 and644, and642′ and644′, respectively, and each having an associated current gain, β_NPN. Here, the common-emitter current gain is less precisely definable, for reasons that will become apparent as operation of the circuit is discussed, but the symbol β_NPNforNPN transistors640 and640′ will be used to mean “base current divided by the sum of emitter currents in a common-emitter configuration” unless otherwise stated. TheNPN transistor640 also includes a base orcontrol electrode646 and a collector orsecond power electrode648, while theNPN transistor640′ includes similarly-numbered elements bearing primes, that is, a base orcontrol electrode646′ and a collector orsecond power electrode648′. Aresistor650 is shown as being connected to theconductor610 and to theemitter642 of theNPN transistor640, andemitter644 is connected directly to theconductor610. TheNPN transistor640′ being analogously coupled to theconductor615 and including theresistor650′ in similar fashion. Specifically, aresistor650′ is connected to theconductor615 and to theemitter642′ of theNPN transistor640′, andemitter644′ is connected directly to theconductor615.
• Theembodiment600 also includes aPNP transistor660 and aPNP transistor660′, each depicted as havingmultiple emitters662 and661, and662′ and661′, respectively, and each having an associated current gain β_PNP. ThePNP transistor660 also includes abase664 and acollector667. Aconductor670, represented as a horizontal central line inFIG. 6, connects thebases664,664′ of the twoPNP transistors660 and660′ together. Thecollector667 of thePNP transistor660 is connected to thebase646 of theNPN transistor640, and both are connected via aresistor668′ to theconductor610. Similarly, thecollector667′ of thePNP transistor660′ is connected to the base646′ of theNPN transistor640′, and both thecollector667′ and the base646′ are coupled via aresistor668 to theconductor615. Theemitter661 ofPNP transistor660 is connected toconductor615 and theemitter661′ ofPNP transistor660′ is connected toconductor610.
• MOSFETs680 and680′ are also included in theembodiment600 ofFIG. 6. TheMOSFETs680,680′ are depicted inFIG. 6 as being enhancement mode n-channel MOSFETs, that is, for example, having a p-type semiconductor body and operating in an enhancement mode. TheMOSFET680 includes a source orfirst power electrode682 that is connected to a node at the juncture of theresistor650 and theemitter642 of theNPN transistor640, and theMOSFET680′ includes a source orfirst power electrode682′ that is connected at the juncture of theresistor650′ and theemitter642′ of theNPN transistor640′.
• TheMOSFET680 includes a gate orcontrol electrode684 and a drain orsecond power electrode686 which are connected together and to theconductor670, with theMOSFET680′ having a gate orcontrol electrode684′ and a drain orsecond power electrode686′ that are connected together and to theconductor670.
• TheMOSFET680 also includes abody electrode688 that is connected to thebase646 of theNPN transistor640, to thecollector667 of thePNP transistor660, to theemitter662′ of thePNP transistor660′, and to theresistor668′. TheMOSFET680′ also includes abody electrode688′ that is connected to the base646′ of theNPN transistor640′, to thecollector667′ of thePNP transistor660′, to theemitter662 of thePNP transistor660, and to theresistor668. EachMOSFET680,680′ has a characteristic threshold voltage V_TH. The threshold voltages V_THfor theMOSFETs680,680′ are assumed to be approximately equal in magnitude and are determinable via factors including doping levels employed during fabrication, as will be explained in more detail with reference toFIG. 8, infra. In alternate embodiments, the threshold voltages V_THforMOSFETs680 and680′ may be formed to be different by, for example, doing additional doping implants.
• In operation, theembodiment600 has a continuum of stable states, including at least two primary or characterizing operating modes. These two primary operation modes are: (i) an OFF mode, corresponding to trace205 or205′ ofFIG. 2, and (ii) an ON mode, corresponding to trace210 or210′ ofFIG. 2.
• In the OFF mode, there is essentially zero volts applied acrossconductors610 and615, and therefore, thetransistors640,640′,660,660′ and680,680′ are all turned OFF, that is, relatively little, to no current is conducted via any of the power electrodes. In some embodiments, in the OFF mode, a relatively small amount of current such as quiescent current or leakage current may flow in or throughcircuit600. The OFF mode may be referred to as a non-conduction mode, low-conduction mode, or low current mode of operation.
• For the ON mode, the operation will be described in terms of a first polarity (for example, positive polarity) of voltage140 (FIG. 1),240 (FIG. 2),340 (FIG. 3) that is presented across theconductors610 and615. As will be described in more detail below, in the first polarity example wherein voltage present on theconductor610 increases relative to a voltage present on theconductor615 to provide a voltage increase with respect to time, the transistors having primed reference characters, that is,transistors660′,680′, and640′, will be turned on and active during this mode of operation. In the second polarity example wherein voltage present on theconductor615 increases relative to a voltage present on theconductor610 to provide a voltage increase with respect to time, the transistors having reference numbers without the primed reference characters, that is,transistors660,680, and640, will be turned on and active during this mode of operation. Accordingly, the circuit inembodiment600 has bidirectional and symmetrical characteristics in that it can respond to electrical transients of either polarity applied acrossconductors610 and615.
• After a “turn-on” voltage is reached, that is, after the magnitude, or the absolute value of, the voltage difference of V_TOis applied acrossconductors610 and615,circuit600 will be in the ON mode of operation, wherein a relatively large amount of electrical current may flow in or throughcircuit600, for example, the amount of electrical current flowing incircuit600 is greater than, and in many cases, substantially greater than, leakage or quiescent current amounts. The ON mode may be referred to as a high-conduction mode or high current mode of operation. Accordingly, the amount of electrical current flowing through thedevice105 is less, and in many cases, substantially less, in the OFF mode compared to the amount of electrical current flowing through thedevice105 in the ON mode. In some embodiments, the amount of electrical current flowing through thedevice105 in the OFF mode is approximately zero amperes, or less than about ten microamperes is some cases, and the amount of electrical current flowing through thedevice105 in the ON mode is greater than approximately 200 milliamperes (mA), and in some cases may range from about 200 milliamperes (mA) to several amperes, depending on the size of elements ofstructure800 described herein.
• As an example, as the voltage presented onconductor610 relative toconductor615 rises towards the positive turn-ON voltage V_TOof thetransient suppression device105, this will initiate some relatively small electrical current flow through theemitter661′-base664′ junction of thePNP transistor660′, and through theresistor668′ and then through theemitter662′-base664′ junction of thePNP transistor660′. As will be described in more detail below, in some embodiments, the current gain oftransistor660′ is relatively small compared to, for example, the current gain ofNPN transistor640′, and as a result most of the current throughPNP transistor660′ is base current. As current begins to flow from the base664′, then electrical current will begin to flow through n-channel MOSFET680′, that is through thedrain686′ and thesource682′ ofMOSFET680′. The relatively small current flow throughPNP transistor660′ andMOSFET680′ may be referred to as “leakage current.” As a result, a voltage onconductor670 will begin to rise toward the threshold voltage ofMOSFET680′, and a voltage V_DSdeveloped across theMOSFET680′, from joineddrain686′ andgate684′, to thesource682′, also begins to increase. This drain-source voltage V_DSwill saturate as it reaches the threshold voltage V_THfor theMOSFET680′. V_THdepends on bias applied to thebody electrode688′ of theMOSFET680′ (that is, V_THincreases with increasing body bias voltage applied tobody electrode688′). The bias on thebody electrode688′ saturates at an effective base-emitter voltage V_BEfor theNPN transistor640′. This base-emitter voltage may be referred to as the turn-ON voltage ofNPN transistor640′.
• There is a relatively small amount of current throughMOSFET680′ until the threshold voltage V_THofMOSFET680′ is attained. After the threshold voltage V_THofMOSFET680′ is attained,MOSFET680′ will turn on more significantly to increase the amount of current throughMOSFET680′. In some embodiments, the threshold voltage V_THofMOSFET680′ may range from about 500 millivolts (mV) to about five volts depending on factors including doping levels and thicknesses of materials employed during fabrication.
• The current through theemitter662′-base664′ junction of thePNP transistor660′ and theemitter661′-base664′ junction of thePNP transistor660′ gives rise to a current through thecollector667′ (I_C=β_PNP*I_B) and the current fromcollector667′ flows to theresistor668. As a result, a voltage will be generated and applied to the base646′ ofNPN transistor640′ and to thebody electrode688′ ofMOSFET680′, and this voltage will rise, resulting in an increase of the threshold voltage V_THofMOSFET680′ since increasing bias voltage applied to thebody electrode688′ acts to increase the effective threshold voltage V_THof theMOSFET680′. The voltage developed across theresistor668 saturates at a voltage that is determined by the voltage drop across the effective base-emitter junction (V_BE) of theNPN transistor640′, which in some embodiments, is a voltage V_BEranging from about 0.6 volts to about 0.7 volts depending on factors including doping levels employed during fabrication.
• A relatively small amount of current is flowing throughcircuit600 at this time prior to the time when the voltage difference acrossconductors610 and615 reaches a value approximating a turn-ON voltage V_TOof thetransient suppression device105, which is the summation of the threshold voltage V_THofMOSFET680′ and the turn-ON voltages V_BEofbipolar transistors660′ and640′. Accordingly, the turn-ON voltage V_TOofcircuit600 ordevice105 may be represented by the equation 2*V_BE+V_TH. In some embodiments, the turn-ON voltage V_TOof thetransient suppression device105 may range from about 1.5 volts to about seven volts.
• Current through theemitter642′ of theNPN transistor640′ also results in a voltage that develops across theemitter resistor650′. That voltage, in turn, tends to reduce current gain A_Iassociated with at least the portion of thecircuit600 that includes theNPN transistor640′. In other words, theresistors650,650′ provide voltage feedback acting to degenerate the effective current gain β of thetransistors640 or640′, respectively, ascurrent I150,250,450 through thetransient suppressor device105 increases, and the reduction in effective transistor gain β as thecircuit600 goes towards the ON mode also tends to stabilize the performance of theembodiment600. The operation and structure ofcircuit600 will be described further below with reference toFIGS. 7,8,9 and10.
• The following description including a plan view inFIG. 7, and cross sectional views, or side views in section, shown inFIGS. 8 and 9, together with a hybrid cross sectional/schematic diagram ofFIG. 10, describes exemplary physical embodiments of the concepts described in the present disclosure.FIG. 7 is a plan view of anembodiment700 of a circuit layout capable of use with thetransient suppression device105 ofFIG. 1. The plan view shown inFIG. 7 finds broad applicability in the relevant arts. The description ofFIGS. 8 and 9 corresponds to a physical embodiment of the schematic circuit diagram600 ofFIG. 6.
• The diagram700 includes anactive region705 that finds application in the context of thetransient suppression device105 ofFIG. 1, the exemplary schematic diagram600 ofFIG. 6 and the view ofFIG. 8, infra. The diagram700 also illustrates contact regions orpads710,715.
• The contact region orpad710 is formed along an interior edge of the active regions and, in this example, surrounded by, theactive region705. Contact region orpad715 is formed along an exterior edge of theactive region705. Thecontact regions710,715 ofFIG. 7 are analogous toconductors510,515 ofFIG. 5, theconductors610,615 ofFIG. 6 and theterminals110,115 ofFIG. 1, for example.
• In the embodiment depicted inFIG. 7, thecontact regions710 and715 may be configured to accept a bond wire coupling theactive region705 to an external circuit that is to be protected from voltage transient effects. In the embodiment ofFIG. 7, theactive region705 is arranged in a roughly square configuration, which might have a surface area comparable to a conventional bond pad, for example, or which may be larger or smaller than such, depending on the desired application that the transient suppression device is intended for. Thecontact regions710 and715 may be formed to be compatible with bump technology, and intended for input/output (I/O) transient protection for an integrated circuit or RF amplifier. In one embodiment, thecontact regions710 and715 have a generally square footprint comprising a contact area on the die of about 400 micrometers on a side, for example.
• In configurations intended for high power applications, thecontact regions710 and715 may be formed using a layered series of conductive materials and/or metals to provide metallurgical compatibility with the various relevant concerns. The squared shapes and serpentine outline shown inFIG. 7 increase the surface area of theactive region705 within the footprint of the overall squared outline, directly increasing the current-carrying capacity of thetransient suppression element105, without substantial increase in overall footprint. These shapes also may provide favorable thermal contact (high thermal conductivity) from thetransient suppression device105, and the substrate from which thetransient suppression device105 is formed, to an external heatsink (not illustrated), if desirable.
• A further desire in many applications is that thetransient suppression device105 provides relatively little OFF-state DC or RF loading of the circuitry to which it is coupled. One facet of this is favorably influenced by designing thetransient suppression device105 to be co-integrable with high-speed circuitry. For example, formingdevice105 integrated with complementary metal oxide semiconductor devices on the same semiconductor substrate may reduce or eliminate parasitic capacitance issues associated with abutting two circuit elements and then joining them with via I/O pads, conductors such as bond wires and the like. Parasitic inductance and RC transmission characteristics and effective switching speeds may be negatively influenced by physical separation and by use of conductive traces, bond wires, pins and/or other inter-element interconnections that may be needed to join elements together.
• Capacitance withintransient suppression devices105 themselves, coupled with need for any significant DC current draw in the OFF state, also limit applications for suchtransient suppression devices105 and tend to compromise performance characteristics realized through inclusion oftransient suppression devices105 in many circuit applications. The examples ofFIGS. 5,6 and8 result in relatively low OFF-state capacitance and thus relatively low OFF-state capacitive loading. In view of the low OFF-state capacitance of these embodiments, a further benefit may be afforded by routing conductors (such as theterminals110 and115 ofFIG. 1) over dielectric areas such as, for example,dielectric regions817 and819 discussed below inFIG. 8, to provide relatively low capacitance between the substrate and theconductors110 and115.
• Accordingly, in one embodiment, relatively large portions of thecontact regions710 and715 are formed over relatively thick dielectric structures that, in turn, are formed to have a relatively low coefficient of permittivity. For example, thecontact regions710 and715 may be formed over dielectric regions (not illustrated inFIG. 7) analogous toregions817 and819 ofFIG. 8, but having broader lateral extent away from edges of the active area. In some embodiments, the dielectric regions may extend from a top surface of the substrate into the substrate approximately 10 microns (μm), but might be formed to have a different depth, such as a depth in a range of between about 3 microns to about 30 microns. In some embodiments,dielectric regions817 and819 may each have a width ranging from approximately five microns to approximately 300 microns. In alternate embodiments, the width ofdielectric regions817 and819 may be about five microns or about ten microns.
• Forming such dielectric regions to have a relative dielectric constant ∈_Rof that of silicon dioxide (for example, about 3.9) or less may be desirable, and in some instances a dielectric structure incorporating sealed voids to provide a relative dielectric constant ∈_Ras low as 1.5 or lower may be desired (for example, seeFIGS. 11,12, and13 for an example of a dielectric structure including sealed voids that may be used fordielectric structures817 and819), or a dielectric region having effective relative dielectric constant ∈_Rof approximately 2.5 may be employed, provided that such dielectric region also conforms with thermal coefficient of expansion, mechanical stability and other concerns applicable in the relevant context.
• Use of such dielectric regions realizes relatively low parasitic capacitance instructure800. As such, the overall RF impedance of the compositetransient suppressor element105 that includes such dielectric regions may be reduced, improving high-frequency performance of thetransient suppressor device105 by leveraging the beneficial OFF-state shunt impedance of, for example, theembodiment600 depicted in schematic form inFIG. 6.
• In some applications, thetransient suppressor device105 may be fashioned as a stand-alone or discrete component that is intended to be attached to other circuitry via a conductive bump. In some applications, one or more of thetransient suppressors105 may be integrated into a more complex circuit, with one or more of thetransient suppressor devices105 being associated with input/output interconnections of the composite integrated circuit to other circuitry external to the die on which the composite device is realized. In such applications, a height (not illustrated) may be used for thecontact regions710 and715 to extend above the substrate and the height may be selected that is compatible with any voltage stand-off requirements associated with elements co-integrated with transient suppressor device(s)105.
• FIG. 8 is a cross section view, or side view, in section, taken along section lines VIII-VIII ofFIG. 7, of astructure800 providing a physical realization of theembodiment600 depicted schematically inFIG. 6.FIG. 8 shows asubstrate802, which in some embodiments, may be asilicon substrate802 that is doped to have a p-plus (p+) or p-conductivity type. In some embodiments, thesubstrate802 is doped to have an acceptor concentration N_Ain a range of from 10¹⁸acceptors/cm³to 10²⁰acceptors/cm³, corresponding to a resistivity of 0.001 Ω-cm to 0.02 Ω-cm, although greater or lesser doping concentrations may be employed. Substrates such assubstrate802 are typically much thicker than structures, regions, or layers such as epitaxial layers that are formed atop, over or as part of the substrate, andsubstrate802 that is heavily doped to be p-type is also compatible with many standard CMOS and RF circuitry processes.
• Thestructure800 ofFIG. 8 includes alayer804 formed over or above thesubstrate802, alayer806 formed atop thelayer804, alayer808 formed overlying thelayer806, and surface dielectric layers orregions807,807′,841, and841′. Thestructure800 also includes acontact region810, acontact region815, isolation structures orregions817 and819, dopedregions820,820′ and a dopedregion821 formed in thelayer806 and which may extend into thelayer808. Thestructure800 also includes aconductive plug823 extending from above a top surface of thelayer808 down to, and providing electrical contact to, the dopedregion821 and thelayer806.Dielectric layers807 and807′ each have an upper or top surface and a lower or bottom surface. In some embodiments, the lower surface ofdielectric layer807 is substantially coplanar to a top surface of dopedregion820 and is substantially coplanar to a top surface ofsemiconductor layer808. The top surface ofdielectric layer807 is substantially coplanar to a top surface ofdielectric structure817, however, in alternate embodiments,dielectric structure817 may be formed so that the top surface ofdielectric structure817 is above or below a plane that is substantially coplanar to the top surface ofdielectric layer807.
• Turning briefly toFIG. 9, a central portion of thestructure800 illustrated inFIG. 8, is shown in an enlarged view inFIG. 9. Referring toFIG. 8 and/orFIG. 9,structure800 also includesdielectric regions813,813′, dopedregions828,828′ and dopedregions830,830′, and includes dopedregions832,832′,834 and834′, each being disposed in theregions820 and820′, respectively. In some embodiments, the dopedregion830 is connected to the dopedregion828; the dopedregion828 is connected to the dopedregion834; the dopedregion828 is in between the dopedregion834 and the dopedregion830; the dopedregion830′ is connected to the dopedregion828′; the dopedregion828′ is connected to the dopedregion834′; and the dopedregion828′ is in between the dopedregion834′ and the dopedregion830′. Further, in some embodiments, wherein a depth of the dopedregion828 is less than a depth of the dopedregion830; a depth of the dopedregion834 is less than a depth of the dopedregion828; a depth of theregion828′ is less than a depth of the dopedregion830′; and a depth of the dopedregion834′ is less than a depth of the dopedregion828′.
• Structure800 also includesdielectric region841 and includesdielectric region856 formed over a portion oflayer808 and over a portion of theisolation structure817. Similarly,structure800 also includesdielectric region841′ and includesdielectric region856′ formed over a portion oflayer808′ and over a portion of theisolation structure819. In addition,structure800 includesdielectric region882 formed over thedielectric region856, and over a portion of thelayer808. Similarly,structure800 includesdielectric region882′ formed over thedielectric region856′, and over a portion of thelayer808′. In some embodiments,FIG. 8 depicts atransient suppression device105 having awidth883 of the active electronic elements or active area (the area between theisolation structures817,819) of about twenty micrometers or less, achieving a relatively compact result consistent with real estate constraints of high density microelectronics devices, such as are employed in realization of communications and computations apparatuses. Accordingly, in these embodiments,layers804 and806 have a width of about twenty microns or less, and layers808 and808′ each have a width of less than ten microns.
• In some embodiments, thelayers804,806 and808 are formed as epitaxial layers of semiconductor material, such as silicon. However, it will be appreciated that doped regions or layers such as804,806 and808 may be realized via other conventional practices, such as ion implantation. In one embodiment, thelayer804 is a p-type epitaxial layer804 having relatively light doping, also known as a mu layer, and is formed atop thesubstrate802. In one embodiment, thelayer804 is formed to have a thickness in a range of about two to about four micrometers (μm), although thicker or thinner layers may be employed. For example, alayer804 having a thickness of about ten micrometers could be employed. Thelayer804 typically incorporates an acceptor concentration N_Ain a range of from 10¹³to about 10¹⁵acceptors/cm³, corresponding to a resistivity of about 4 Ω-cm.
• In some embodiments, thelayer806 is an n-type layer806, and is formed over thelayer804 by ion implantation into the surface of themu layer804 and prior to any further epitaxial layer growth, to have a thickness of on the order of about one and one-half micrometers to about two micrometers, although other thicknesses may be usefully employed. Alternatively, the n-type layer806 may be formed subsequent to formation of epitaxial layers above the p-type layer804, by high energy ion implantation. In either case, the n-type layer806 may be more heavily doped than theunderlying mu layer804 in some embodiments. In one embodiment, thelayer806 is formed to have a donor concentration N_Din a range of from about 5×10¹⁸donors/cm³to about 5×10²⁰donors/cm³.
• In some embodiments, thelayer808 is an n-type epitaxial layer808, and is formed is formed over thelayer806. In some embodiments, thelayer808 has a thickness of about one micrometer. In some embodiments, thelayer808 is formed to have a donor concentration N_Dof about 2×10¹⁶donors/cm³, corresponding to a resistivity of on the order of about 0.3 Ω-cm. Theregions820,820′ are doped to provide p-type areas, and also are referred to as “p-tubs.”
• Layers orregions802,804,806, and808 may each be referred to as a semiconductor material in some embodiments. It will be appreciated that all of theregions802,804,806 and808 may formed using other doping concentrations and thickness and that other combinations of doping and thickness may be used. In addition, in some embodiments, p-type regions802 and804 may be optional. For example, in embodiments whereindevice105 is a discrete device that will be bump attached,epitaxial layer808 could be formed on a heavily doped n-type substrate. In these embodiments, contact todevice105 would be throughcontact regions810 and815 and there would be no contact or attaching to the back of the die as the n-type substrate would be floating. In embodiments wherein the die includes more than one transient suppressor device, then structure800 ofFIG. 8, having p-type substrate802 and p-type region804, may be used and also structure800 may be used whendevice105 is integrated with other types of devices.
• Contact structures orregions810 and815 are formed over the ensemble oflayers804,806 and808 comprising thestructure800. Contactregions810 and815 are respectively analogous toterminals110,115 ofFIG. 1,conductors610,615 ofFIG. 6, contact orpad regions710,715 ofFIG. 7, andconductors510,515 ofFIG. 5. Thecontact regions810,815, may also be referred to as “ohmic” contacts, and provide relatively low resistance bidirectional electrical contact to selected portions of thetransient suppressor device105 ofFIG. 1, for example. In one embodiment, thecontact region810 may extend over theisolation structure819 and thecontact region815 may extend over theisolation structure817.
• Isolation structures817,819 provide lateral electrical isolation of thestructures804,806 and808, and may provide edge termination mitigating effects of depletion region curvature, and also provide capacitive isolation of theconductive contacts810 and815 from thesubstrate802, among other things. In some embodiments, thedielectric structure817 is between substantially all of theinterconnect815 and thesemiconductor substrate802 to reduce parasitic capacitance between theinterconnect815 and thesemiconductor substrate802, and similarly, thedielectric structure819 is between substantially all of theinterconnect810 and thesemiconductor substrate802 to reduce parasitic capacitance between theinterconnect810 and thesemiconductor substrate802. In some embodiments, thedielectric structure817 is between at least a portion of, or a majority of, theinterconnect815 and thesemiconductor substrate802 to reduce capacitance between theinterconnect815 and thesemiconductor substrate802.
• In some embodiments, theisolation structures817 and819 may be formed to have a depth extending through the epitaxial orcounter-doped layers804,806 and808 and extending into thesubstrate802, as illustrated inFIG. 8, or may be formed to have a different depth in other embodiments. It is also noted that in some embodiments,dielectric structures817 and819 may be optional, that is,structure800 may be formed withoutdielectric structures817 and819. In other words, in some embodiments, other isolation means may be used to isolate the active region ofstructure800. Further, in some embodiments,dielectric structures817 and819 may be separate discrete structures formed substantially simultaneously, or formed at different times, an in other embodiments,dielectric structures817 and819 may be a unitary structure that surrounds the active region ofstructure800.
• Thestructure800 also includes aregion821, which may be an n-type dopedregion821 adjacent and electrically coupled to a heavily n-type dopedpolycrystalline silicon plug823 formed in an opening or trench above theregion821. Theregion821 and theplug823 collectively form a low-resistance electrical contact to thelayer806. An n-type dopedregion828 is in the p-tub region820, and an analogous n-type dopedregion828′ is formed in the doped p-tub region820′.
• Referring toFIGS. 8 and 9,structure800 includes conductivesidewall gate portions811 and811′,conductive portions812 and812′,dielectric structures814 and814′, dielectric layer orregion816, anddielectric layers818 and818′,dielectric layers841 and841′, anddielectric regions882,882′, and884.
• Referring toFIGS. 5,6,8, and9,regions806,808, and808′ correspond toconductor570 ofFIG. 5 and toconductor670 ofFIG. 6. That is,conductor670 ofFIG. 6 corresponds to n-type buriedlayer806 and n-type epitaxial layers808 and808′ which together form a common contact, node, or bus, withindevice105, that is attached to the collectors ofNPN transistors640′ and640, the gates and drains of n-channel MOSFETs680 and680′, and the base electrodes ofPNP transistors660 and660′.
• Referring toFIGS. 6,8, and9, the n-type region828 corresponds to theemitter644′ of theNPN transistor640′ ofFIG. 6. Similarly, the n-type region828′ corresponds to theemitter644 of theNPN transistor640 ofFIG. 6. The portion of the p-type region820 underlying theregion828 corresponds to the base646′ of theNPN transistor640′. Similarly, the portion of the p-type region820′ underlying theregion828′ corresponds to thebase646 of theNPN transistor640. The portion of the n-type layer808 and the n-type layer806 underlying the n-plus (n+)region828 corresponds to thecollector648′ of theNPN transistor640′. Similarly, the portion of the n-type layer808′ and the n-type layer806′ underlying the n-plus (n+)region828′ corresponds to thecollector648 of theNPN transistor640.
• Thestructure800 further includes heavily p-type doped or p-plus (p+)regions830 and830′ disposed in the p-type regions820 and820′, respectively. In addition,structure800 includes n-type,doped regions832 and834 formed in p-type region820 and includes n-type,doped regions832′ and834′ formed in p-type region820′. As will be discussed below, in some embodiments, n-type dopedregions832,832′,834 and834′ may be more lightly doped than n-type dopedregions828 and828′. Further,doped regions832,832′,834 and834′ may be formed during the same ion implant step, anddoped regions832 and832′ may be floating regions, and therefore, are essentially inactive during the operation ofdevice105.
• Structure800 may be referred to as a merged structure. In the context of the disclosed structure, a merged structure refers to a semiconductor device which has regions that independently can perform as different elements of a schematic representation of the device. In other words, thestructure800 ofFIG. 8 can be represented or correspond tocircuit600 ofFIG. 6, and regions or elements ofstructure800 can correspond to multiple elements ofcircuit600 so that regions or elements ofstructure800 perform multiple functions ofcircuit600. For example, the n-type dopedregion834 corresponds to thesource682′ of theMOSFET680′ and the n-type dopedregion834′ corresponds to thesource682 of theMOSFET680. In addition, the n-type dopedregion834 also corresponds to theresistor650′ and to theemitter642′ of theNPN transistor640′. Similarly, the n-type dopedregion834′ also corresponds to theresistor650 and to theemitter642 of theNPN transistor640.
• Thedielectric layer841 corresponds to the gate dielectric or gate oxide layer for theMOSFET680′.Conductive gate structure811 overlies thedielectric layer841 and functions as thegate electrode684′ ofMOSFET680′ and that portion of the p-type region820 under thedielectric layer841 functions as the channel region for theMOSFET680′. Similarly, thedielectric layer841′ corresponds to the gate dielectric or gate oxide layer for theMOSFET680.Conductive gate structure811′ overlies thedielectric layer841 and functions as thegate electrode684 ofMOSFET680 and that portion of the p-type region820′ under thedielectric layer841′ functions as the channel region for theMOSFET680. The portion ofregion808 adjacent to the channel region ofMOSFET680′ corresponds to thedrain686′ ofMOSFET680′ and the portion ofregion808′ adjacent to the channel region ofMOSFET680 corresponds to thedrain686 ofMOSFET680.
• Region830′ and the portion of p-type region820′ under p-type region830′ corresponds to theemitter661′ ofPNP transistor660′; the base664′ ofPNP transistor660′ corresponds to portions oflayers808′,806, and808, and thecollector667′ ofPNP transistor660′ corresponds to the portion of p-type region820 under p-type region830. Similarly,region830 and the portion of p-type region820 under p-type region830 corresponds to theemitter661 ofPNP transistor660; thebase664 ofPNP transistor660 corresponds to portions oflayers808,806, and808′, and thecollector667 ofPNP transistor660 corresponds to the portion of p-type region820′ under p-type region830′. As will be described below with reference toFIG. 10,resistor668 is a pinch resistor formed in the portion ofregion820 underregions828 and834, and similarly resistor668′ is a pinch resistor formed in the portion ofregion820′ underregions828′ and834′. Theemitter662 ofPNP transistor660 is in series withresistor668 and theemitter662 corresponds to a portion ofregion820 that is under a portion ofregion828 and/or a portion ofregion834, and similarly, theemitter662′ ofPNP transistor660′ is in series withresistor668′ and theemitter662′ corresponds to a portion ofregion820′ that is under a portion ofregion828′ and/or a portion ofregion834′.
• In some embodiments, theconductors810 and815 may be formed using an initial layer of titanium having a thickness of about 600 Angstroms to form a titanium silicide surface layer with the surfaces of the dopedregions828,830,828′ and830′ of the embodiment ofFIG. 8 of thetransient suppression device105 and this layer of titanium may be overlaid with a layer of titanium nitride having a thickness of about 800 Angstroms. Titanium is one of many materials useful in forming high electrical conductivity or “ohmic” contacts to silicon devices and may also be formed atop silicon dioxide or other materials employed in thetransient suppression device105. It will be appreciated that many different selections of materials may be used in this capacity and that a variety of formation techniques may be useful. As an example, in some embodiments, a layer of titanium sufficiently thick to provide the desired thicknesses may be annealed in dry nitrogen, for example via rapid thermal annealing, to form a strongly-bonded, conductive titanium nitride layer that also is capable of functioning as a chemical and metallurgical barrier.
• Thecontact regions810 and815 may be further formed using, for example, a layer of 14,000 Angstroms of suitable metals or alloys, such as, for example, aluminum, aluminum-copper alloy, tungsten or any other suitable metal formed over the metallurgical barrier layer and that conform to the outlines of the active region705 (FIG. 7). Other materials or metals may be employed, to provide suitable electrical and thermal conductivity, physically and metallurgically robust contacts, and surfaces capable of interconnection to circuitry external to thesubstrate802 from which thetransient suppressor device105 is formed over, when such interconnection is anticipated.
• FIG. 10 is a composite physical and schematic diagram1000 of a portion ofstructure800. The composite diagram1000 corresponds to several blended or distributed elements of themerged structure800, including the multiple-emitter NPN transistor640′ ofFIG. 6, portions of the multiple-emitter PNP transistors660 and660′ ofFIG. 6, theresistors650′ and668, and theMOSFET680′.
• As discussed above with reference toFIGS. 8 and 9,resistor650′ corresponds to n-type region834, which is more lightly doped than n-type region828.Resistor650′ extends from the end ofregion828 that is connected toregion834, thoughregion834 to the end of the region of834 that corresponds to thesource682′ ofMOSFET680′ and is adjacent to the channel region ofMOSFET680′.
• As is illustrated in composite diagram1000,resistor668 is a pinch resistor formed in the portion ofregion820 underregions828 and834, that is,resistor668 is formed in the region of p-type region820 that is pinched between n-type regions828 and834 and underlying n-type region808 so that n-type regions are on both sides of the p-body region820 to formpinch resistor668. As is illustrated in composite diagram1000,resistor668 is formed inregion820, extending from the end ofregion830 that is connected toregion828 to the end ofregion834 that is next to the channel region ofMOSFET680′.Pinch resistor668 is further illustrated by showing adiode1003 that represents the p-n junction between theregions820 and828 and adiode1004 that represents the p-n junction between theregions820 and808.Resistor668 is defined by the amount of p-type dopant that is inregion820, and as is discussed below, theregion820 is formed through a triple implant or chain implant process. In addition, this implant process may be used to establish the threshold voltages ofMOSFETs680 and680′, the forward voltage drops or turn-on voltages V_BEand current gains of the bipolar transistors ofcircuit600.
• Referring toFIGS. 6,8,9, and10,resistor668 is a resistor that may be used to prevent latch-up by providing a current path to remove charge from the base-emitter junction ofNPN transistor640′ so that when a transient voltage applied acrossconductors610 and615 falls, the base drive current available to the base646′ ofNPN transistor640′ is reduced so thatNPN transistor640′ will not remain on. Latch-up is a condition wherein the only way to turn off the devices ofcircuit600 may be by removing the applied drive voltages tocircuit600 which may be undesirable in many applications.
• In the OFF state, internal portions of thecircuit600 provide relatively high impedances, which can result in internal charge buildup or accumulation, when a relatively large voltage has been applied to the power electrodes. As a result, fluctuations in the voltage applied to the power electrodes can lead to “false triggering” ofcircuit600. One way of reducing probability of false triggering is to include bleed resistors, that is, resistors that permit stored charge to bleed away from areas in which it can contribute to such false triggering.
• One function that is fulfilled via theresistors668 and668′ of theembodiment600 ofFIG. 6, and by the resistors558 and558′ of theembodiment500 ofFIG. 5, is to facilitate such bleeding of stored charge. Accordingly, these embodiments of transient suppression devices, acting in the role of thetransient suppression device105 ofFIG. 1, are less vulnerable to false triggering than some examples from the prior art.
• Resistors650 and650′ provide temperature compensation incircuit600. Theresistors650 and650′ of thecircuit600 ofFIG. 6 provide a thermal ballast incircuit600 when thetransient suppression device105 experiences an electrical stress that is sufficient to result in significant temperature excursion.Resistors650 and650′ have positive temperature coefficients, meaning that as the temperature rises, the resistance ofresistors650 and650′ increases. On the other hand,MOSFETs680 and680′ each have a negative temperature coefficient meaning that as the temperature rises, the threshold voltages V_THofMOSFETs680 and680′ decrease. Similarly,bipolar transistors640,640′,660, and660′ each have a negative temperature coefficient, meaning that as the temperature rises, the forward base-emitter voltage drop or turn-on voltages V_BEoftransistors640,640′,660, and660′ decrease. Therefore, by havingresistors650 and650′ present, the clamp voltage ofcircuit600 may be regulated to a better degree than a circuit that does not includeresistors650 and650′. As energy is dissipated withinstructure800, the temperature ofstructure800 will rise, and therefore, the turn-on voltage V_TO, or clamp voltage, ofstructure800, would be reduced as more energy is dissipated due to the negative temperature coefficients oftransistors640,640′,660,660′,680, and680′.Resistors650 and650′, having positive temperature coefficients, will at least partially counteract or counterbalance the effect of the negative temperature coefficients oftransistors640,640′,660,660′,680, and680′, to provide thermal compensation or temperature compensation incircuit600 to reduce variations in the clamp voltage ofcircuit800 due to temperature excursions.
• Emitter642 is a distributed type emitter since towards the center of structure800 (that is, toward an area including conductive plug823) the gain of the emitter decreases since there is more degeneration. In other words, as discussed above,resistor650 is also formed as part ofregion834′ which is the region whereemitter642 is formed. Accordingly, toward the center ofstructure800, the gain ofemitter642 is reduced and the resistance ofresistor650 increases.
• WhenNPN transistor640′ is turned ON, current through theemitter642′ of theNPN transistor640′ also results in a voltage that develops across theresistor650′. The voltage across theresistor650′, in turn, tends to reduce current gain A_Iassociated with at least the portion of thecircuit600 that includes theNPN transistor640′. In other words, theresistors650,650′ provide voltage feedback acting to degenerate the effective current gain β of thetransistors640 or640′, respectively, as current (150,250,350) through thetransient suppressor device105 increases, and the reduction in effective transistor gain β as thecircuit600 goes towards the ON mode also tends to stabilize the performance of thecircuit600.
• As discussed above, structure800 (FIGS. 8,9,10) is a merge structure, wherein regions, or portions of regions, ofstructure800 may perform more than one function, or correspond to more than one element, of circuit600 (FIG. 6). For example, as discussed herein, p-type region820 performs more than one function, in that, portions ofregion820 serve as a portion ofresistor668, as the base646′ ofNPN transistor640′, as thecollector667′ ofPNP transistor660′, as theemitter662 ofNPN transistor660, as a portion of theemitter661 of thePNP transistor660, and as the channel ofMOSFET680′.
• Referring again toFIGS. 6,8,9, and10, the operation ofcircuit600 in view ofstructure800 will be explained with the following example beginning with the application of a voltage present onconductor810 increasing relative to a voltage present on theconductor815. As the voltage onconductor810 rises with respect to the voltage onconductor815, electrical current flow is initiated throughresistor668′, which is formed withinregion820′ as a pinch resistor. The current flows fromregion830′ toregion820′, and as discussed herein, these regions correspond toemitters661′ and662′ ofPNP transistor660′.
• Emitter661′ is a distributed emitter in that it is distributed throughoutregion820′, however, the portion ofregion820′ that is belowregion830′ will be the lowest resistance portion, and therefore, will be the most active portion ofemitter661′ meaning that when the current starts to flow fromregion830′ a majority of the current will flow through the portion ofregion820′ belowregion830′. The current will then flow intoregion808′,region806, and thenregion808, wherein these regions correspond to the base664′ ofPNP transistor660′ and theconductor670. As is illustrated instructure800,PNP transistor660′ has a relatively wide base width sincebase664′ corresponds toregions808′,806, and808, and therefore,PNP transistor660′ may be a relatively low gain transistor since the heavily doped n-type buriedlayer806 has a relatively large amount of charge in this region. In other words,PNP transistor660′ has a relatively low current gain since thebase regions808′,806, and808 have a relatively large amount of charge due to the relatively high doping concentration inregion806. With current flow in the emitter-base region ofPNP transistor660′, this will give rise to current flow into thecollector667′ ofPNP transistor660′ which corresponds to p-type region820. Accordingly, thecollector667′ and theemitters661′ and662′ ofPNP transistor660′ are on opposite sides ofpolysilicon plug823. Similarly, thecollector667 and theemitters661 and662 ofPNP transistor660 are on opposite sides ofpolysilicon plug823.
• With current flow inregions806 and808 (corresponding to the base664′ ofPNP transistor660′ and the conductor670), this will initiate current flow through n-channel MOSFET680′. That is, current will flow between thesource682′ (corresponding to region834) and thedrain686′ (corresponding to a portion of region808) ofMOSFET680′. This current will be relatively low until the threshold voltage V_THofMOSFET680′ is reached.
• The current fromcollector667′ will flow to resistor668 (corresponding to a portion of region820), which will result in an increasing voltage applied to the base646′ (corresponding to a portion of region820) ofNPN transistor640′ and thebody electrode688′ (corresponding to a portion of region820), which will result in an increase of the threshold voltage V_THofMOSFET680′ untilNPN transistor640′ turns on. As may be appreciated,PNP transistor660′ andNPN transistor646′ provide a four-layer silicon controlled rectifier (SCR) structure, which is a four-layer PNPN structure wherein the emitter-base PN junction ofPNP transistor660′ corresponds to one PN junction and the collector-base PN junction ofNPN transistor640′ corresponds to the other PN junction.
• If the voltage difference across conductors610 (corresponding to conductor810) and615 (corresponding to conductor815) reaches a value approximating a turn-ON voltage V_TOof thetransient suppression device105, which is the summation of the threshold voltage V_THofMOSFET680′ and the turn-ON voltages V_BEofbipolar transistors660′ and640′ (2*V_BE+V_TH), then the voltage acrossconductors610 and615 clamps at a voltage of about V_TOand then significantly more current can flow throughcircuit600. This mode of operation may be referred to as the ON mode or a high current mode of operation, and in this mode of operation,transistor640, having emitters644 (corresponding toregion828′) and642 (corresponding to a portion ofregion834′), base646 (corresponding to a portion ofregion820′), and collector648 (corresponding to portions ofregions808′,806 and808), operates in an inverse mode or reverse mode of operation, wherein the collector ofNPN transistor640 operates as the emitter ofNPN transistor640, and wherein theemitters642 and644 ofNPN transistor640 operate as a collector ofNPN transistor640 to conduct current after a voltage difference of V_TOis applied acrossconductors610 and615.
• WithNPN transistor640 operating in a reverse mode of operation to conduct current, this results in a five-layer NPNPN structure incircuit600, which can handle relatively more current than a four-layer SCR structure since it has an n-type region at both ends of the five-layer structure. The five-layer NPNPN structure corresponds to n-type region828′ (layer1), p-type region820′ (layer2), combined n-type regions808′,806, and808 (layer3), p-type region820 (layer4), and n-type region828 (layer5). Therefore, when the voltage difference of V_TOis applied acrossconductors610 and615 andcircuit600 is in a high current mode of operation, a majority of the current through thecircuit600 passes throughPNP transistor660′, through theNPN transistor640′ and through theNPN transistor640 which is operating in a reverse mode. Further, with such a five-layer structure, the predominant current throughstructure800 is electrons. Switching speed, particularly the turn-OFF speed, and peak current-carrying capacity are often direct functions of the mobility of the charge-carrying elements, that is, electrons or holes. Since electrons have 2.3 times the mobility compared to holes, that means thestructure800 can for the same size handle 2.3 times more current, compared to a ESD structure that is primarily based on hole-conduction.
• The switching speed, and thus the turn-ON and turn-OFF speeds, are related to the carrier diffusion velocity for sweeping charge carriers out of the relevant regions in thetransient suppression device105. Thus, when a voltage surge suppression device is formed that is based on hole-conduction mechanisms and carrier diffusion, the turn-OFF speed suffers by a factor of 2.3 in comparison to analogous electron-conduction devices, particularly in devices relying on diffusion and/or minority-carrier phenomena in depletion regions. As a result, both area requirements and switching speeds are strongly related to charge-carrier mobility and both are favorably influenced via usage of higher-mobility charge carriers, such as electrons, in at least some types of applications. In some embodiments, the response time, or turn-on and turn-off times, may be about one nanosecond or less. In other words, referring briefly back toFIGS. 3 and 4, the turn-on time (T₁to T₂) of thedevice105 is less than about one nanosecond and the turn-off time (T_1′ to T_2′) of the voltage clamp circuit is less than about one nanosecond in some embodiments.
• In some embodiments, the current gain oftransistor660′ is less than the current gain oftransistor640′. Similarly, to maintainstructure800 as a symmetrical device, the gain oftransistor660 is less than the current gain oftransistor640. To avoid issues with latching or snapback, and to promote rapid switching and relatively low effective ON mode resistance (the inverse of the slope, dI/dV, of theline segments210,210′ ofFIG. 2), the product of the current gains of theNPN transistor640′ and thePNP transistor660′ is less than, or near or slightly less than unity, that is, β_NPN*β_PNP<1. Similarly, in these embodiments, the product of the current gains of theNPN transistor640 and thePNP transistor660 is near or slightly less than unity.
• Current gain β of a bipolar transistor is a function of the amount of charge that is stored in the base region, and numerous other design variables, as is understood in the relevant arts. Doping levels are parameters amenable to adjustment as a portion of the fabrication process, without requiring re-tooling of mask sets and the like. Structuring the current gain product to be less or slightly less than unity also provides latitude with respect to fabrication variations and for temperature excursions.
• Accordingly, the overall current gain, or current gain product, oftransistors660′ and640′ ofcircuit600 may be set by establishing the doping levels instructure800 so that the current gain product is less than one. For example, if the current gain oftransistor640′ is about 100, then the current gain of transistor may be set to be about 0.009 so that the current gain product is slightly less than one, that is, about 0.9 in this example. If the current gain product is greater than one,structure800 may go into a latch mode in response to a voltage transient applied acrosscontact regions810 and815. Ifstructure800 goes into a latch mode, or if snapback occurs, then it may be relatively difficult to returncircuit600 to an OFF mode or resetcircuit600. For example, turning the transistors ofcircuit600 off may require removing the applied signals or voltages that are applied tocircuit600. By setting the current gain product oftransistors640′ and660′ to be less than one, this will allowcircuit600 to reset, or return or transistor to an OFF mode from an ON mode without having to remove the signal, or whatever voltage difference, is applied across theconductors610 and615. With the current gain product oftransistors640′ and660′ less than one, this will enablecircuit600 to operate as illustrated inFIGS. 2,3, and4. For example,circuit600 may operate as follows: in response to a voltage transient applied acrossconductors610 and615, as the voltage difference rises to a voltage level greater than the turn-on V_TOor clamp voltage ofcircuit600, this voltage will be clamped at the turn-on voltage V_TO, then after the voltage transient drops to a level below the clamp voltage thecircuit600 with return to an OFF mode.
• Accordingly, the structures, devices, and circuits disclosed herein, for example,device105,circuits500 and600, andstructure800, provide in some embodiments a device or circuit configured to arrest electrical transient events responsive to a voltage excursion manifested between the first and second terminals, wherein the circuit provides a characteristic voltage-current response having symmetry about a predetermined voltage of less than ten volts, the circuit being configured to provide a positive low ON-mode DC effective resistance, to provide little OFF-mode loading of ancillary elements, presenting an absence of distortion of modulation bearing information that is impressed upon a carrier wave coupled to the circuit, and to exhibit capacity for switching from the ON mode to an OFF mode responsive to an a voltage excursion resulting from the modulation manifesting a time duration within a range of voltages relevant to such switching of less than one nanosecond. In other embodiments, the structures, devices, and circuits disclosed herein, for example,device105,circuits500 and600, andstructure800, provide a clamp circuit formed on a semiconductive substrate and having first and second conductors, the clamp circuit being configured to provide a clamp voltage of about seven volts or less with respect to a voltage difference between the first and second conductors, the clamp circuit including clamping characteristics that are symmetric about a predetermined voltage, wherein the clamp circuit is configured to provide a current-voltage characteristic exhibiting only positive effective direct current resistance at the first and second conductors.
• Further, in some embodiments, an electrical device is disclosed, wherein the electrical device comprises a bidirectional polarity, voltage transient protection device such asdevice105 described herein. The voltage transient protection device may include a first terminal and a second terminal and may have at least two modes of operation comprising an OFF mode and an ON mode, wherein the amount of electrical current flowing through the voltage transient protection device is less in the OFF mode compared to the amount of electrical current flowing through the voltage transient protection device in the ON mode. In some embodiments, The voltage transient protection device may include a PNP bipolar transistor such astransistor660 or660′ having a turn-on voltage of V_BE1and a first current gain, β₁, and include a NPN bipolar transistor such astransistor640 or640′ having a turn-on voltage of V_BE2and a second current gain, β₂, wherein the product of the current gains of the PNP transistor and the NPN transistor is less than unity, that is, β₁*β₂<1. In some embodiments, the current gain of the PNP bipolar transistor, β₁, is less than the current gain of the NPN bipolar transistor, β₂.
• The transient protection device may further include a field effect transistor (FET) such asMOSFET680 or680′ having a threshold voltage of V_TH. The voltage transient protection device may transition from the OFF mode to the ON mode if a voltage difference approximately equal to, or greater than, a predetermined turn-on voltage such as V_TOis applied across the first and second terminals of the voltage transient protection device and the voltage transient protection device may transition from the ON mode to the OFF mode if a voltage difference less than the predetermined turn-on voltage such as V_TOis applied across the first and second terminals of the voltage transient protection device. In some embodiments, the predetermined turn-on voltage (V_TO) is approximately equal to the sum of V_BE1, V_BE2, and V_TH, that is, V_TO≅V_BE1+V_BE2+V_TH, wherein the magnitude of the predetermined turn-on voltage is symmetric about a predetermined reference voltage such as, for example, a reference voltage of zero volts.
• In the following portion of the disclosure, exemplary processes for realizing thestructure800 ofFIGS. 8 through 10, respectively, are described with reference toFIGS. 11 through 28. Those of skill in the art will appreciate that variations from the parameters and materials described herein are also possible, and within the scope and spirit of the present disclosure.
• FIGS. 11 to 13 are used to describe at least one embodiment ofdielectric structure817, and a method of makingdielectric structure817.FIG. 11 is a cross section view of a portion ofstructure800 at one stage during manufacturing in accordance with an embodiment of the present disclosure.Structure800 includes p-type substrate802 and p-type layer804 described in the discussion ofFIG. 8 above.
• Adielectric layer960 overlieslayer804. In an embodiment of the wafer process,dielectric layer960 comprises silicon dioxide (SiO₂). The layer of SiO₂is thermally grownoverlying layer804 having a thickness ranging from approximately 500 Angstroms to approximately 5000 Angstroms (Å). Amasking layer961 is formed overlyingdielectric layer960. Maskinglayer961 is patterned exposing portions ofdielectric layer960. The exposed portions ofdielectric layer960 are removed revealing theunderlying epitaxial layer804. Maskinglayer961 is then removed. An etching process is then performed to form, for example, a matrix of hexagonal vertical hollow wells, openings, orcavities965. In particular, an anisotropic etching process is used to etch substantially vertically through at least theepitaxial layer804 and, preferably, at least part way intosubstrate802. In this embodiment,vertical cavities965 are approximately 2.0 microns wide and spaced 0.4 microns apart from one another and define a matrix of vertically extending structures or walls. Using the anisotropic etching process,vertical cavities965 are etched throughepitaxial layer804 and intosubstrate802 to a depth ranging from approximately 3 microns to approximately 30 microns. The etching ofvertical cavities965 creates silicon pillars orwalls966 between thecavities965. Theinnermost wall966ais adjacent to an outer portion of theactive area995 ofdevice105. Theactive area995 may also be referred to as an active region.Silicon walls966 are approximately 0.4 microns wide.Dielectric layer960 is affected by the above wafer process steps such thatdielectric layer960 is reduced in thickness, in some embodiments, from a thickness of about 5000 Å to approximately 3000 Å.
• An optional process act may be performed that removes material fromsilicon walls966. For example, in some embodiments, a silicon etch is performed that etches exposed portions ofsilicon walls966,epitaxial layer804, andsubstrate802. In these embodiments of the wafer process, the silicon etch thinssilicon walls966 to a width or thickness of approximately 0.2 μm.
• Referring toFIG. 12, a thermal oxidation process is performed that forms silicon dioxide on any exposed silicon area. In particular, the silicon ofsilicon walls966 ofFIG. 11 may be substantially completely converted to silicon dioxide formingsilicon dioxide walls976 in the form of a matrix of vertically extending dielectric structures. In other words, the silicon between the sidewalls ofwalls966 may be substantially converted to silicon dioxide. In addition, as shown inFIG. 12, during the thermal oxidation process, the exposed silicon surface of the innermost wall (labeled966ainFIG. 11), the bottom of cavities965 (labeled962 inFIG. 11) and the outermost wall (labeled966binFIG. 11) are likewise converted tothermal oxide material976a,972 and976b, respectively. Depending on the application, it may be desirable to form, for example, deposit, further dielectric material to increase the thickness of the thermal oxide dielectric material. A further consideration is the time required to form the dielectric material and stress applied to the structure. For example, an additional deposition of apolysilicon layer970 is performed. Then, a thermal oxidation step oxidizes thepolysilicon layer970 forming a dielectric layer that increases the amount of dielectric material onsilicon dioxide walls976,976a,976band thebottom portions972 ofcavities965. After these thermal oxidation processes are performed, the silicon dioxide material formed on the sidewalls ofopenings965 may be represented as a homogeneous dielectric silicon dioxide material in subsequent figures, wherein the portions ofdielectric structure817 betweenopenings965 are labeled976 in subsequent figures, the portion ofdielectric structure817 nearest theactive area995 is labeled976a, the bottom portions ofdielectric structure817 is labeled976, and the portion ofdielectric structure817 furthest from theactive area995 is labeled976b. In addition, the portion ofpolysilicon layer970 formed overlying layer215 is converted to silicon dioxide and may be illustrated as a homogeneous dielectric structure over p-type layer804 that is labeled960 in subsequent figures. In other words, the thickness oflayer960 may be increased through this additional optional polysilicon deposition and thermal oxidation processes in some embodiments.
• Referring toFIG. 13, a dielectric material is applied to the die. In some embodiments of the wafer process, a low-pressure deposition of TEOS (tetra-ethyl-ortho-silicate)oxide980 is applied at the surface ofstructure800. Some of the deposited material builds up in each opening ofvertical cavities965 gradually reducing the size of the opening until the opening is closed forming a dielectric plug or layer in the upper portions ofcavities965. The remaining lower portions ofcavities965 are not filled in this embodiment. In alternate embodiments, the lower portions of thecavities965 could be filled with a dielectric material, such as a material comprising oxide or nitride, if so desired, to form a solid, filleddielectric structure817. Note that a continuous layer of dielectric material is formed in eachcavity965 by way ofdielectric layer980 anddielectric portions976,976a,976b, and972. This structure of dielectric material, includingair gaps965, is denoted asdielectric structure817. In some embodiments of the wafer process, approximately 11,000 Å ofTEOS980 is deposited such that an upper region ofvertical cavities965 are sealed to form a plurality of sealedair gaps965. A thermal oxidation process may follow that densifies the TEOS that is part ofdielectric structure817.Sealed air gaps965 may also be referred to as a plurality of closed cells, wherein the closed cells may include theair gaps965, the verticalsilicon dioxide structures976, and thedielectric structures980.
• In some embodiments, an oxide chemical mechanical planarization (CMP) step is then performed to planarize the oxide over the top surface ofstructure800 after the dielectric material deposition. The CMP step removes from the top surface portions ofTEOS layer980 anddielectric layer960 and creates a substantiallyplanar surface975 over the top surface. It should be noted that althoughvertical cavities965 are sealed at the upper surface bydielectric layer980,vertical cavities965 are not filled with solid material and comprise a substantial amount of empty space or air gaps. Aprotective layer990 is then applied overlying the oxide on the top surface. In some embodiments of the wafer process, a layer ofsilicon nitride990 approximately 500 Å thick overliesplanar surface975. As mentioned previously, an alternate process flow that does not require an oxide CMP step could be developed should CMP not be desired or available. The surface should be sufficiently planar to prevent step coverage problems with subsequent wafer processing steps.
• In some embodiments,dielectric structure817 may be referred to as a dielectric platform or dielectric region, and may be formed greater than 10 microns wide and greater than 3 microns deep. Passive devices and/or electrical interconnects, electrodes, or contacts may be formed overlyingdielectric structure817. In some embodiments,dielectric structure817 is formed to a depth greater than 4 microns. Moreover,dielectric structure817 can be formed at these dimensions or greater without significant stress being added to the die. Also, it should be understood that various different manufacturing processes can be employed to form the dielectric structure. For example, thecavities965 may be filled forming asolid dielectric structure817. In addition, in some embodiments, dielectric structure819 (FIG. 8) may be formed at the same time that dielectricstructure817 is formed, and dielectric structure819 (FIG. 8) may be formed using similar, or the same, processes used to formdielectric structure817.
• As mentioned previously, optional polysilicon deposition and thermal oxidation processes may be performed to increase the thickness of the dielectric material ofdielectric structure817. In some embodiments, prior to formingdielectric capping layer980, polysilicon is deposited intovertical cavities965 forming a polysilicon layer on the bottom and sidewalls. For example, 1000 Å of polysilicon may be deposited intovertical cavities965. The polysilicon may then be oxidized to form a 2200 Å thermal oxide layer invertical cavities965. A second, 1000 Å of polysilicon may then be deposited and oxidized to form a second 2200 Å oxide layer invertical cavities965. The combination forms a 4400 Å oxide layer invertical cavities965. Other techniques known to one skilled in the art can also be applied that increase the amount of dielectric material. For capping thecavities965, the upper portions ofcavities965 should not be made so large that they cannot be closed by a process step such as the low pressure TEOS deposition.
• In general, thedielectric structure817 is a non-conductive structure having a relatively low dielectric constant. From a structural perspective, the oxide formed on the bottom portions ofcavities965 and the sidewall965aadjacent to theactive area995 should not be formed to a thickness where stress is induced into the substrate that produces warpage or dislocation defects in the wafer. Thus, thedielectric structure817 is designed to reduce stress imparted to the wafer when the dielectric structure comprises a substantial portion of the die area.
• Thedielectric structure817 may serve as a support structure that has sufficient structural strength to allow the formation of interconnect, passive components, or active devices over thedielectric structure817. To achieve this, in some embodiments, vertical support structures such aswalls976 are formed that support a top surface layer such aslayer980. The vertical support structures and top surface layer comprise a dielectric material. In one embodiment, empty compartments underlying the top surface layer are formed between the vertical support structures to form air gaps that lower the dielectric constant of the dielectric platform. Conversely, a solid or filleddielectric structure817 could be formed which would have a relatively higher dielectric constant if desired. In some embodiments,dielectric structure817 may be an array of hexagonal cells having vertical walls formed of silicon dioxide when viewed looking down on the top surface. The center region of each hexagonal cell is an empty void or space. A cap or top surface layer is formed to seal each hexagonal cell. The diameter of a cell indielectric structure817 may be determined by the capping process. The diameter of the cell may be selected to allow the build up of deposited dielectric material near the opening near the top surface which closes off and seals the cell without filling the cell up (with the deposited dielectric material such as TEOS). Similar spacing constraints would apply to other air gap dielectric structures that utilize a capping process.
• Thedielectric structure817 may also reduce parasitic capacitances of devices formed using the wafer, thereby extending the frequency response of the devices. Thedielectric structure817 separates conductive regions from one another thus a low dielectric constant is preferred to minimize the capacitance. The lowest dielectric constant for a dielectric structure may be achieved by maximizing the volume of empty space in the dielectric structure between conductive regions which form the parasitic capacitance. In particular, the number of cells indielectric structure817 or the area of the die thatdielectric structure817 comprises is related to reducing the parasitic capacitances.
• FIG. 14 is a cross section view ofstructure800 at a subsequent stage in the manufacturing or the processing ofstructure800.Structure800 includessubstrate802 andlayers804,806 and808, described in the discussion ofFIG. 8 above, whereinlayers804,806, and808 are prepared to provide a substantially planar top surface, which supports alayer807.FIG. 14 also showsdielectric structures817 and819, wherein embodiments ofdielectric structures817 and819 are described above.
• In some embodiments, alayer807 is formed as an initial or pad layer of silicon dioxide having a thickness ranging from about 600 to about 700 Angstroms via thermal oxidation of a portion of the silicon material at the top surface oflayer808. In some embodiments, the thickness ofthermal oxide layer807 is about 670 Angstroms.
• Structure800 includeslayers854 and854′, providing chemical selectivity with respect to thelayer807, formed on thelayer807 and extending overdielectric structures817 and819 as shown inFIG. 14. In some embodiments, thelayers854 and854′ may be formed from Si₃N₄(silicon nitride) by conventional plasma techniques to have a thickness ranging from about 250 Angstroms to about 1,500 Angstroms.Layers856 and856′ are formed onlayers854 and854′, respectively. In some embodiments, thelayers854 and854′ may be formed from Si₃N₄(silicon nitride) to have a thickness of about 250 Angstroms, by conventional plasma techniques, and thelayers856 and856′ may be formed of silicon dioxide via conventional TEOS (tetra ethyl ortho silicate) deposition techniques, to have a thickness of about 11,000 Angstroms. For example, a layer of Si₃N₄may be deposited using a chemical vapor deposition (CVD) process, then a layer of TEOS may be deposited on the Si₃N₄layer, and then these layers may be pattered using photolithography and etching processes to form anopening864 andlayers854,854′,856, and856′. The photolithography and etching may include forming a layer of a radiation-sensitive material, such as photoresist (not shown), over the deposited TEOS layer, then exposing the photoresist using UV radiation to form a mask, and then etching through the TEOS layer and the Si₃N₄layer using the photoresist mask, stopping at or inoxide layer807 to form opening864 andlayers854,854′,856, and856′.Opening864 extends down to, but not penetrating, thelayer807, and dividing thelayers854,854′,856,856′. In some embodiments, theopening864 has a width of about 1.7 micrometers, for example. Photolithography processes or operations may sometimes be referred to as masking operations or acts.
• Thelayers807,854,854′,856, and856′, provide protection of the top surface oflayer808 from chemical contamination during fabrication, and also provide selective etching capabilities such that thelayers854,854′,856, and856′ may be etched without substantially etching thelayer807 and vice versa.
• FIG. 15 is a cross section view ofstructure800 at a subsequent stage in the manufacturing or the processing ofstructure800.FIG. 15 shows alayer814, which may be formed from Si₃N₄atop the resultant structure shown inFIG. 14, and within the opening864 (FIG. 14) to have a thickness of about 1500 Angstroms, and formed by conventional plasma techniques, for example. In various embodiments, the thickness oflayer814 may range from about 1000 Angstroms to about 1500 Angstroms.
• Alayer853 is formed atop thelayer814. In some embodiments, thelayer853 may comprise polycrystalline silicon or polysilicon, and may have a thickness ranging from about 3,000 Angstroms to about 5,000 Angstroms. As a result, anopening863 is formed within the opening864 (FIG. 14), wherein opening863 is narrower than theopening864 and is self-aligned to theopening864. Theopening863 may have a width that is less than that of theopening864 by approximately twice the thickness of thelayer853.
• FIG. 16 is a cross section view ofstructure800 at a subsequent stage in processing.FIG. 16 shows a result of conventional anisotropic etching of thelayer853, to form anopening865 that is similar in size to the opening863 (FIG. 15), but extending to thelayer814, and which is masked within the opening by residual portions of thelayer853. These residual portions may be referred to assidewall spacers853.
• FIG. 17 is a cross section view ofstructure800 at a subsequent stage in processing.FIG. 17 shows that another anisotropic etch operation is performed to extend opening865 through the layer814 (FIG. 16) and 807 (FIG. 16) and partially into theunderlying layer808. In one embodiment, the silicon nitride layer814 (FIG. 16) is etched via conventional anisotropic plasma etching, such that exposed portions of thelayer814 atop thelayers856 and856′ have been stripped, to provide residual components orportions814 and814′ shown inFIG. 17. The portion of the layer807 (FIG. 16) exposed within theopening865 is then etched to provideresidual portions807 and807′ shown inFIG. 17, and a portion of the material of thelayer808 exposed via etching of thelayer807 is also etched. Conventional etching techniques for silicon nitride and silicon dioxide provide a chemical selectivity of about 5:1 with respect to etching of polycrystalline silicon, and, as a result,residual portions853 of the polycrystallinesilicon sidewall spacers853 ofFIG. 16 are also etched to reduce the height ofsidewall spacers853.
• FIG. 18 is a cross section view ofstructure800 at a subsequent stage in processing.FIG. 18 shows that a conventional anisotropic etch oflayer808 and some of thelayer806 has been performed through opening ortrench865, which has a width and nearly vertical sidewalls to extend completely through thelayer808 and stopping at or in buried n-type layer806. During this etch, the residual portions orsidewall spacers853 ofFIG. 17 that had been formed during the acts described above with reference toFIGS. 15-17 are completely eroded. As a result, a narrow and relatively vertically-walled slot865 is realized, which is self-aligned with respect to structures formed with reference to theopening864 ofFIG. 14.
• Following the initial etch described with reference toFIG. 18, in some embodiments, an optional oxidation step and a subsequent anisotropic etch can provide silicon dioxide sidewalls813 and813′ within thetrench865. The subsequent anisotropic etch will remove oxide at the bottom oftrench865 to exposesurface806 as shown inFIG. 18 so that the bottom oftrench865 is void of dielectric. Silicon dioxide sidewalls813 and813′ function as barriers to migration of materials and/or defects such as dislocations from materials within thetrench865 to regions adjacent thetrench865. Other conventional processes also may be used to providesidewalls813 and813′ that provide migration barriers. In some embodiments, thedielectric sidewalls813 and813′ range from about 100 Angstroms to several hundred Angstroms in thickness. Conventional anisotropic etching of silicon dioxide provides high chemical selectivity such that underlying or adjacent silicon nitride portions are not substantially etched and thus does not require particularly tight process control.
• The optional barrier formation act to form silicon dioxide sidewalls813 and813′, in turn, results in containment of dopant introduced later in processing, thus avoiding incurring increased dislocation density and the like by inhibiting propagation of such defects from material that is later deposited into thetrench865. In some embodiments, thetrench865 has a width on the order of about 4,000 Angstroms. The processes described result in thetrench865 being self-aligned respective to other features formed as a portion of thetransient suppression device105.
• FIG. 19 is a cross section view ofstructure800 at a subsequent stage in processing.FIG. 19 shows a dopedregion821 beneath a tip of the bottom of thetrench865, and alayer823 that has been deposited to a thickness of about 8,000 Angstroms. In some embodiments, thelayer823 comprises polycrystalline silicon, and may be doped or undoped as deposited. Dopant may be introduced, for example by ion implantation of arsenic or antimony at a dose of about 3×10¹⁵ions/cm³, to dope the material forming the layer823 n-type via subsequent heat treatment or anneal process, which also causes a portion of dopant from thelayer823 to dope theregion821 as well thereby increasing the n-type doping level in the proximity of the contact region ofpolysilicon region823 tolayer806. As a result, the material forming thelayer823 is conductive and is electrically coupled to thelayer806 via an ohmic-type interconnection.
• Arsenic diffuses less rapidly than many other types of n-type dopants in monocrystalline silicon such as, for example, phosphorus, and thus presents a relatively robust and well-defined location for the implanted species through subsequent thermal cycling associated with subsequent processing acts, and also tends to segregate at grain boundaries in polycrystalline silicon during subsequent heat treatments. In either case, when there are nosidewalls813 and813′ forming a barrier within thetrench865, a doped region is formed in the semiconductive material, that is, layers808 and808′, that surround portions of thelayer823. In other words,layers813 and813′ serve as a way of preventing the dopant oflayer823 from doping regions of808 and808′ adjacent to layer823.
• FIG. 20 is a cross section view ofstructure800 at a subsequent stage in processing.FIG. 20 shows a result of an anisotropic etch of thelayer823 to provide aconductive plug823 of conductive material such as, for example, doped polycrystalline silicon. As a result of the etching, the portions oflayer823 that are over the upper surfaces oflayers856 and856′, and over the upper portions of814 and814′, are removed.
• FIG. 21 is a cross section view ofstructure800 at a subsequent stage in processing.FIG. 21 shows a result of a masking operation followed by an isotropic wet oxide etch of the materials orlayers856 and856′. In some embodiments, the isotropic etch is a slight over-etching of thelayers856 and856′, that is, is performed for a period of time sufficient to remove exposed portions of thelayers856 and856′, and then includes enough additional time to ensure complete clearing of the portions ofoxide layers856 and856′ from exposed areas, but stopping at the nitride layers854 and854′ because of chemical selectivity.
• A chain implant has been performed to provide three p-type doping regions that will becomeregion820 after subsequent thermal processing. The upper region may be referred to as a surface implant, the middle region may be referred to as a central implant, and the lower region may be referred to as a lower implant.
• For the surface, central, and lower implants corresponding to theresultant region820, the dose employed for the surface implant is used to establish the threshold voltage V_THof theMOSFET680′ ofFIG. 6. In particular, the dose employed for the upper region or surface implant is linearly related to the threshold voltage V_THof theMOSFET680′ ofFIG. 6. As a result, adjustment of the dose of the implant used to provide the surface implant also provides a measure of control for the voltage at which thetransient suppression device105 displays nonlinear or voltage clamping behavior, without requiring modification to photomasks used to manufacture thetransient suppression devices105. The doses employed for the central implant and the lower implant corresponding to resultant p-type region820 affect the current gain β_NPNof theNPN transistor640′ and also affect the resistance of theresistor668.
• Similarly, for the surface, central, and lower implants corresponding to theresultant region820′, the dose employed for the surface implant is used to establish the threshold voltage V_THof theMOSFET680 ofFIG. 6. In particular, the dose employed for the upper region or surface implant is linearly related to the threshold voltage V_THof theMOSFET680 ofFIG. 6. The doses employed for the central implant and the lower implant corresponding to resultant p-type region820′ affect the current gain β_NPNof theNPN transistor640 and also affect the resistance of theresistor668′.
• A chain implant may be achieved by programming an implanter to do a series or chain of implants at different energies and doses. The higher the energy, the deeper the penetration for the implant. In some embodiments, the surface implants may be formed by a boron implant to a dose of about 1×10¹³ions/cm², at an energy of about 60 keV, the central implants may be formed by a boron implant to a dose of about 2×10¹³ions/cm²at an energy of about 90 keV, and the lower implants may be formed by a boron implant to a dose of about 2×10¹³ions/cm², at an energy ranging from about 110 keV to about 120 keV.
• The surface, central, and lower implants may be subsequently activated to form p-type regions820 and820′ by some subsequent thermal processing such as, for example, a rapid thermal anneal (RTA) process. These implants will establish a doping profile that is reflected by their energies and dopant concentrations, and accordingly, this is a way of profiling and establishing characteristics that will becomedoped regions820 and820′ out of three implants.
• FIG. 22 is a cross section view ofstructure800 at a subsequent stage in processing.FIG. 22 shows a result of an isotropic wet nitride etch of thematerials854 and854′ and814 and814′. In some embodiments, the etching is done using hot phosphoric acid and is timed to remove an amount of silicon nitride ranging from about 750 Angstroms to about 1250 Angstroms, and results in thinning of sidewalls formed ofmaterial814 and814′ by that amount.
• FIG. 23 is a cross section view ofstructure800 at a subsequent stage in processing. A thermal treatment act is performed to anneal thestructure800 and activate the species implanted during the chain implant as described above with reference toFIG. 21, so that the three implants of the chain implant essentially merge together to form dopedregions820 and820′. In addition, some of the processes described below with reference toFIG. 23, are thermal processes such as, for example, the thermal oxidation processes to formlayers805,805′,816,818,818′,841, and841′, and these thermal processes will also activate the ion-implanted dopants implanted during the chain implant as described above with reference toFIG. 21 to form dopedregions820 and820′.
• An isotropic wet oxide etch is used to etch thematerial807 and807′, and this etch will also etch thematerials856 and856′. In some embodiments, the etchedmaterials807 and807′ comprise silicon dioxide and the etching is performed using hydrofluoric acid that is timed to remove an amount of oxide ranging from about 750 Angstroms to about 1250 Angstroms.
• Layers ofmaterial841 and841′ are formed. In some embodiments, thematerials841 and841′ comprise silicon dioxide formed via thermal oxidation of the silicon of thelayers808 and808′. During this thermal oxidation act, the upper exposed surface of theplug material823 is also oxidized to form a layer ofsilicon dioxide816 at the upper exposed surface of theplug material823. In some embodiments, thematerials841 and841′ are grown to a thickness ranging from about 100 to about 400 Angstroms, and may be grown to a thickness of about 200 Angstroms. Thematerials841 and841′ ultimately form the gate dielectric material orgate oxide841 and841′ for theMOSFETs680′ and680 ofFIG. 6, respectively. As a result, the thicknesses of thematerials841 and841′ are linearly related to the threshold voltages V_THofMOSFETs680′ and680, respectively. Accordingly, control of the thicknesses of thematerials841 and841′ provides another mode for adjustment of the threshold voltages V_THofMOSFETs680′ and680 which does not require alterations of masking tooling. A subsequent oxidization act (“gate re-oxide”) may be performed to increase the thickness of thematerials841 and841′.
• A blanket layer of electrically conductive material is formed after thedielectric materials841 and841′ have been formed. In some embodiments, a layer of polycrystalline silicon is deposited to a thickness of about 3,000 Angstroms and is then anisotropically etched, leavingpolysilicon sidewalls811 and811′. In some embodiments, the layer of polysilicon deposited may range in thickness from about 2,000 Angstroms to about 5,000 Angstroms and the anisotropic etching is carried out as an “over-etch” act, that is, the time duration of the etch exceeds the time needed to substantially clear polysilicon from areas in which it forms horizontal sheets, by a factor of 1.1 or 1.2 (ten to twenty percent over etch). In some embodiments, thematerial forming sidewalls811 and811′ is n-type polycrystalline silicon doped with phosphorous as deposited. As discussed above,sidewalls gate portion811 functions as thegate684′ ofFET680′. Accordingly, in some embodiments, thegate684′ of theFET680′ comprises an electrically conductive material such as polysilicon and thegate684′ corresponding to sidewallgate811 illustrated inFIG. 8 is formed so that the gate length of theFET680′ is approximately equal to a deposition thickness of the layer of electricallyconductive material811. Accordingly, in the embodiment discussed with reference toFIG. 23, the gate length of theMOSFETs680 and680′ are determined non-photolithographically. The deposition of material such as polysilicon can be controlled with great accuracy in a wafer fabrication facility (wafer fab). The gate length ofMOSFET680′ is determined by the thickness of depositedpolysilicon layer811. What this means is that a transistor can be produced with a state of the art gate length (for example, 0.2-0.3 microns or lower) in a wafer fab having photolithographic capabilities greater than 0.35 microns. The short channel length of the transistor may result in extended frequency response. Thedevice105 may be built at relatively lower cost since production cost is directly related to the photolithographic capability of the wafer fab. Moreover, tighter control over gate lengths may be achieved with lower variance because of the control wafer processing facilities have over material deposition thicknesses (such as polysilicon). Although the gate length ofMOSFETs680 and680′ are discussed as being determine non-photolithographically, the methods and apparatuses described herein are not limited in this regard. For example, in alternate embodiments, the gate lengths ofMOSFETs680 and680′ may be determined by the minimum feature size of the available photolithography equipment.
• A subsequent thermal oxidation process is then performed that continues to form a layer ofsilicon dioxide816 at the upper exposed surface of theplug material823, and also forms layers ofsilicon dioxide805 and805′ over the exposed surfaces of thepolysilicon sidewalls811 and811′, respectively. Further, this subsequent thermal oxidation process also thickens the silicon dioxide on p-type regions820 and820′ to form silicon dioxide layers818 and818′ that are thicker thanlayers841 and841′.
• FIG. 24 is a cross section view ofstructure800 at a subsequent stage in processing.FIG. 24 shows results of an ion implantation to provide dopant that forms n-type dopedregions834 and834′. In some embodiments, the implantation includes implanting phosphorous or arsenic at a dose of about 2×10¹⁴ions/cm², at an energy of about 60 keV. Portions of the dopedregions834 and834′ respectively correspond tosources682′ and682 ofMOSFETs680′ and680 ofFIG. 6, to theemitters642′ and642 of theNPN transistors640′ and640, and to theresistors650′ and650.
• FIG. 25 is a cross section view ofstructure800 at a subsequent stage in processing. In some embodiments, an etch is performed using hot phosphoric acid, to remove portions ofmaterial814 and814′, leaving gaps (not shown inFIG. 25) between thematerials811 and823 and between thematerials811′ and823. This etch may also etch a relatively small amount oflayers854 and854′. In some embodiments, this etch may be timed to remove an amount of nitride ranging from about 2,000 Angstroms to about 3,000 Angstroms.
• A layer of electrically conductive material such as, for example, polycrystalline silicon, is deposited and then is anisotropically etched, resulting insidewall structures812 and812′ on portions ofdielectric layers805 and805′ and also formingconductive portions812 and812′ in the gaps between thematerials811 and823, and between thematerials811′ and823, to form a “bridge” or coupling structure to electrically couple electricallyconductive materials811 and823 and811′ and823, whereinconductive materials811,811′,812,812′, and823 form a homogeneous or contiguous electrically conductive structure or material such that aportion823 extends to dopedregion821 and aportion811 is overdielectric layer841. Thesidewall structures812 and812′ on the portions oflayers805 and805′ are floating structures that are electrically isolated from other electrically conductive materials ofstructure800. In some embodiments, thematerials812 and812′ comprise phosphorous doped n-type polycrystalline silicon having a thickness of about 500 Angstroms and the anisotropic etch of the polycrystalline silicon may be timed to remove an amount of polysilicon ranging from about 750 Angstroms to about 1,250 Angstroms. In other embodiments, the deposited polycrystalline silicon may be undoped polycrystalline silicon, and subsequent thermal steps may dope the undoped polycrystalline silicon from the dopant that is present inmaterials811,811′, and823.
• FIG. 26 is a cross section view ofstructure800 at a subsequent stage in processing.FIG. 26 shows results of formation of a layer of dielectric material that is then pattered using photolithography and etching processes to formstructures882,882′, and884 and to formopenings883 and883′. In one embodiment, thematerial forming structures882,882′, and884 is silicon dioxide deposited to a thickness of about 11,000 Angstroms via a conventional TEOS process. In some embodiments, edges of thematerials882,882′, and884 are “feathered” to enhance step coverage for later-deposited conductive materials, using conventional isotropic and anisotropic etching techniques. The processes to formstructures882,882′, and884 may be relatively low temperature processes so that these processes have relatively little effect onregions834 and834′ at this stage, so that the processes used to formstructures882,882′, and884 do not cause diffusion ofregions834 and834′.
• FIG. 27 is a cross section view ofstructure800 at a subsequent stage in processing.FIG. 27 shows results of two sequential masking acts. The initial masking act defines areas that are then ion implanted to providedoped regions830 and830′. The subsequent masking act defines areas that are then ion implanted to providedoped regions828 and828′. More specifically, the initial masking act covers the central region ofstructure800, proximate to structure884, and exposes the portions ofopenings883 and883′ outside of the central region. Then, a boron implant is performed at a dose ranging from about 3×10¹⁴ions/cm²to about 3×10¹⁵ions/cm², and at an energy of about 50 keV, to formregions830 and830′. In some embodiments, the does of the boron implant is at about 1×10¹⁵ions/cm². After the boron implant, the subsequent masking act is performed to cover the outer region ofstructure800 and to expose the portions ofopenings883 and883′ proximate to structure884, that is, to cover the regions whereregions830 and830′ are formed. Then, an arsenic implant is performed at a dose ranging from about 2×10¹⁵ions/cm²to about 1×10¹⁶ions/cm², and at an energy of about 75 keV, to formregions828 and828′. In some embodiments, the does of the arsenic implant is at about 3×10¹⁵ions/cm².
• Regions832 and832′ result from the formation ofregions828,828′,830, and830′.Regions832 and832′ are floating regions and are essentially inactive during the operation ofdevice105. As may be appreciated,doped regions832,832′,834 and834′ are formed from the same ion implant step discussed with reference toFIG. 24. As discussed above, in some embodiments, n-type dopedregions832,832′,834 and834′ may be more lightly doped than n-type dopedregions828 and828′.
• FIG. 28 is a cross section view ofstructure800 at a subsequent stage in processing.FIG. 28 shows results of heat treatment ofstructure800, for example via conventional rapid thermal annealing (RTA), that diffuses and activates the species implanted as described with respect toFIG. 27, providingdoped regions828,828′,830, and830′. Referring back toFIG. 8, after the RTA operation is performed as described with reference toFIG. 25,conductive layers810 and815 may be formed by depositing a layer of electrically conductive material such as, for example, a metal, and then patterning this layer of electrically conductive material using photolithography and etching processes to formcontact regions810 and815.
• As mentioned above,transient suppression device105 may be integrated with metal oxide semiconductor (MOS) devices on the same semiconductor substrate. As an example,FIG. 29 is a cross section view ofstructure800 integrated with a n-channel MOSFET900. Accordingly, together,structure800 andMOSFET900 form an integrated circuit (IC), wherein the integrated circuit comprises electricaltransient suppression device105 andMOSFET900. For purposes of clarity, all the elements ofstructure800 andMOSFET900 are not shown. The n-channel MOSFET900 includes a n-type source region910 in a portion of p-type layer804, a n-type drain region920 in a portion of p-type layer804, agate oxide930 over a portion oflayer804, at least a portion of agate940 overgate oxide930, and achannel region950 formed in a portion oflayer804 undergate oxide930, and betweenregions910 and920. In this example, n-channel MOSFET900 is isolated from the active area ofdevice105 bydielectric structure817.
• FIGS. 30 to 34 illustrate another embodiment for forming thegates684 and684′ ofFETs680 and680′, respectively. The example method for forminggates684 and684′ shown inFIGS. 30 to 34 is an alternative method compared to the method of forminggates684 and684′ shown inFIGS. 23 to 25.
• FIG. 30 is a cross section view of thestructure800 shown inFIG. 22 at a subsequent stage of manufacturing in accordance with an embodiment of the present disclosure. A thermal treatment act is performed to anneal thestructure800 and activate the species implanted during the chain implant as described above with reference toFIG. 21, so that the three implants of the chain implant essentially merge together to form dopedregions820 and820′.
• An isotropic wet oxide etch is used to etch thematerial807 and807′, and this etch will also etch thematerials856 and856′. In some embodiments, the etchedmaterials807 and807′ comprise silicon dioxide and the etching is performed using a timed hydrofluoric acid etch.
• Layers ofmaterial841 and841′ are formed. In some embodiments, thematerials841 and841′ comprise silicon dioxide formed via thermal oxidation of the silicon of thelayers808 and808′. During this thermal oxidation act, the upper exposed surface of theplug material823 is also oxidized to form a layer ofsilicon dioxide816 at the upper exposed surface of theplug material823. In some embodiments, thematerials841 and841′ are grown to a thickness ranging from about 100 to about 400 Angstroms, and may be grown to a thickness of about 200 Angstroms. Thematerials841 and841′ ultimately form the gate dielectric material orgate oxide841 and841′ for theMOSFETs680′ and680 ofFIG. 6, respectively. As a result, the thicknesses of thematerials841 and841′ are linearly related to the threshold voltages V_THofMOSFETs680′ and680, respectively. Accordingly, control of the thicknesses of thematerials841 and841′ provides another mode for adjustment of the threshold voltages V_THofMOSFETs680′ and680 which does not require alterations of masking tooling.
• A blanket layer of electrically conductive material is formed after thedielectric materials841 and841′ have been formed. In some embodiments, a layer ofpolycrystalline silicon891 is conformally deposited to a thickness of about 1000 Angstroms. In addition, a layer of photoresist is spun on and baked over the layer ofpolycrystalline silicon891, and then portions of this layer of photoresist are removed to form photoresist layers892 and892′ and to expose portions of the layer ofpolycrystalline silicon891.
• FIG. 31 is a cross section view ofstructure800 at a subsequent stage in processing. The layer ofpolycrystalline silicon891 is etched to remove portions ofpolycrystalline silicon891, leavingpolysilicon layers891 and891′.
• FIG. 32 is a cross section view ofstructure800 at a subsequent stage in processing.FIG. 32 shows the result of removing thephotoresist892 and892′ and a result of an anisotropic wet nitride etch of thematerials814 and814′. In some embodiments, the etching ofmaterials814 and814′ is done using a timed hot phosphoric acid etch to form gaps between thematerials891 and823 and between thematerials891′ and823.
• FIG. 33 is a cross section view ofstructure800 at a subsequent stage in processing. Another layer of electrically conductive material such as, for example, polycrystalline silicon having a thickness of about 2000 Angstroms, is deposited and then is etched, resulting insidewall structures811 and811′ on portions ofpolysilicon materials891 and891′.Conductive polysilicon materials811 and811′ also have portions in the gaps between thematerials891 and823, and between thematerials891′ and823 to form a “bridge” or coupling structure to electrically couple electricallyconductive materials891 and823 and891′ and823, whereinconductive materials891,891′,811,811′, and823 form a homogeneous or contiguous electrically conductive structure or material such that aportion823 extends to dopedregion821,portions811 and891 are overgate oxide layer841, andportions811′ and891′ are overgate oxide layer841′.
• In some embodiments, thematerials811,811′,891, and891′ comprise phosphorous doped n-type polycrystalline silicon. In other embodiments,materials891 and891′ may be undoped polysilicon as deposited, and then may be doped by subsequent implant acts.
• A subsequent thermal oxidation process is then performed that continues to thicken the layer ofsilicon dioxide816 at the upper exposed surface of theplug material823, and also forms layers ofsilicon dioxide805 and805′ over the exposed surfaces of thepolysilicon materials811,811′,891, and891′, respectively. Further, this subsequent thermal oxidation process also forms the silicon dioxide layers818 and818′ on p-type regions820 and820′, respectively. In some embodiments,layers818 and818′ may be thicker thanlayers841 and841′, respectively.
• FIG. 34 is a cross section view ofstructure800 at a subsequent stage in processing.FIG. 34 shows the result of an ion implantation to provide dopant that forms n-type dopedregions834 and834′. In some embodiments, the implantation includes implanting phosphorous or arsenic at a dose of about 2×10¹⁴ions/cm², at an energy of about 60 keV. Portions of the dopedregions834 and834′ respectively correspond tosources682′ and682 ofMOSFETs680′ and680 ofFIG. 6, to theemitters642′ and642 of theNPN transistors640′ and640, and to theresistors650′ and650. In some embodiments, the structure shown inFIG. 34 may be further processed using processes similar to those processes described above with reference toFIGS. 26 to 28.
• In the preceding description, characteristics of transient suppression devices were described, and examples illustrating operation, circuit realization and fabrication aspects were discussed. In the following description, several applications for which the transient suppressor of the present disclosure is suited are presented. More particularly, applications in this section include a standalone component, an integrated circuit and an RF amplifier.
• FIG. 35 is a cross section view of a stand-alone,discrete component2600 embodiment incorporating the transient suppression device of the present disclosure.FIG. 35 shows asubstrate2601 for an electronic device, optionaldielectric packaging material2602, and atransient suppression device2605 formed in thematerial2602.Transient suppression device2605 may have similar structure, or the same structure, as thetransient suppression device105 described above.Transient suppression device2605 includes a terminal2610 and a terminal2615 (analogous to first andsecond terminals110,115 ofFIG. 1, for example) that are coupled to interconnections orinterconnects2626 and2628. Interconnections2626 and2628 are coupled toconductors2620 and2625, respectively, whereinconductors2620 and2625 are formed on thesubstrate2601.
• In some embodiments, thesubstrate2601 may be a ceramic or composite dielectric material such as materials conventionally employed in formation of hybrid circuits, or may be a printed circuit board (PCB) or other basal member. The first andsecond interconnections2626 and2628 may be bump interconnections forming conductive interconnections that may also provide mechanical stability, and may be formed using metals or metallic components that are melted to secure the various components together, or may be formed using conductive polymers such as conventional conductive epoxies.Bump Interconnects2626 and2628 are external interconnects forelectrically coupling terminals2610 and2615 to interconnects or circuitry external to thediscrete component2600.
• In some embodiments, when intended for such standalone applications, thetransient suppression device2605 may be formed using a silicon substrate that is heavily doped to be n-type. In other words, theexample structure800 ofFIG. 8 was shown as employing ap+ type substrate802 and lightly p-type dopedmu layer804. That embodiment addresses facilitation of formation of the transient suppression device, for example, together with some other types of integrated circuit elements, for example CMOS or bipolar transistors. That embodiment also includes capacitive isolation between various circuit elements, but these portions of that embodiment may not be needed in some applications. As discussed above, in some embodiments, p-type regions802 and804 may be optional. For example, in embodiments whereindevice105 is a discrete device that will be bump attached,epitaxial layer808 could be formed on a heavily doped n-type substrate. In these embodiments, contact todevice105 would be throughcontact regions810 and815 and there would be no contact or attaching to the back of the die as the n-type substrate would be floating.
• In the example ofFIG. 35, when a voltage stress begins to manifest across theconductors2620 and2625, thetransient suppression device2605 begins to conduct, as described above with respect toFIGS. 2 through 10. As a result, one or more voltage-sensitive components that are also coupled to theconductors2620 and2625 are only exposed to voltage excursions limited to a predetermined range, with the result that an electronic assembly incorporating thetransient suppression device2605 is relatively more robust with respect to electrical disturbances compared to other transient suppression devices. While the example ofFIG. 35 is illustrated in flip-chip form, it will be appreciated that other techniques for coupling thetransient suppression device2605 to a circuit element for which electrical stress protection is desirable may also be employed, such as pins that couple to sockets, zero insertion force or ZIF interconnections, bond wires and the like.
• FIG. 36 is a simplified plan view of anintegrated circuit embodiment2700 incorporating one or more of the transient suppression devices of the present disclosure.FIG. 36 shows asubstrate2702 including semiconductive material, portions including one or moretransient suppression devices2705 and2705′,interconnection sites2706, some of which may incorporate atransient suppression device2705, aportion2707 incorporating one or more transistors or electronic switching or amplification devices and aninterconnection region2709.
• Theportion2707 may comprise digital circuitry, analog circuitry or both. At one or moreselected points2711 within theportion2707 that otherwise could present liability or fragility with respect to transient electrical stresses, atransient suppression device2705′ may be incorporated. Theperipheral interconnection region2709 often presents particularized vulnerability to environmentally-related electrical voltage stresses or excursions, for example due to electrostatic discharge encountered in handling or mounting of theintegrated circuit2700, or causing electrical disturbance at an exterior connection point connected to theperipheral interconnection portion2709. For at least this reason, one or moretransient suppression devices2705 may be incorporated in theinterconnection region2709.
• When signals exchanged between theintegrated circuit2700 and electrical apparatus external to theintegrated circuit2700 are required to be high frequency RF signals or to have sharply-delineated time characteristics, such as rising or falling edges of high speed digital signals, capacitive or other RF loading of the interconnections is an important concern. In these types of applications, such loading may be reduced by forming the interconnections atop structures providing dielectric isolation of the interconnections from conductive portions of theintegrated circuit2700. As examples, theisolation structures817 and819 ofFIG. 8 provide increased electrical isolation of theinterconnections810 and815 from circuit elements formed usingsubstrate802 and also increased electrical isolation from thesubstrate802 itself, in comparison to many other approaches. Theisolation structures817 and819 may be formed to have a relative dielectric constant ∈_Rof that of silicon dioxide or to have a relative dielectric constant ∈_Rthat is substantially less than that of silicon dioxide, and may be formed to have a depth extending into the substrate which may be several micrometers or which may be ten or more micrometers.
• FIG. 37 is a simplified schematic diagram of aRF amplifier2800 embodiment including at least one transient suppression device of the present disclosure.FIG. 37 depicts theamplifier2800 to include atransient suppression device2805 having first andsecond conductors2810 and2815 that, in turn, are coupled to an input port2830 (represented in part by a dashed vertical line).Transient suppression device2805 may have similar structure, or the same structure, as thetransient suppression device105 described above. Theinput port2830 couples anRF input signal2832 RF_INto theamplifier2800. Theamplifier2800 also includes an output port2835 (represented in part by a dashed vertical line) that couples anoutput signal2839 RF_OUTto other circuitry (not shown). AFET2890 having a first power electrode orsource2892 and a second power electrode ordrain2894 and a control orgate electrode2896 is shown as being coupled in a common-source amplifier configuration, with thesource2892 being coupled to theconductor2825 and thegate2896 being coupled to thesecond conductor2820. Thedrain2894 is coupled through asuitable circuit Z_D2897 to terminal2898 to receive a power supply voltage V_DDand is also coupled to theoutput port2835.
• When theamplifier2800 is intended to provide gain together with high linearity, such as a class A or a class AB amplifier, it is important that thetransient suppression device2805 not introduce clipping or otherwise distort theinput signal2832 RF_IN. One reason that a symmetric IV curve, such as shown inFIG. 2, is desirable in some applications for atransient suppression device2805 is that the origin of the IV curve can be made to be coincident with the average of theinput signal2832 RF_IN, with the voltages V_TOand −V_TOfalling outside of a region of maximum excursion for theinput signal2832 RF_IN.
• Improved apparatuses, processes and techniques have been disclosed for electrical stress protection elements that may be utilized as a discrete component or co-integrated with other types of circuitry. The disclosed electrical stress protection devices have the capacity for repeated electrical stress relief that provide for relatively rapid and effective abatement of electrical stresses, appropriate peak current-carrying capacities, relatively high switching ON and OFF speeds, automatic capability for resetting to the OFF state when appropriate, bidirectional conduction properties, relative immunity or reduced sensitivity to false triggering, threshold voltages that are programmable during manufacturing that are realized via relatively few processing operations, relatively less electrical loading of other electrical circuitry that may result in reduced DC and RF signal loading characteristics, and threshold characteristics tailorable to specific criteria during manufacturing without necessarily requiring modification of tooling.
• It will be appreciated that while the disclosure describes transient suppressor devices in the context of silicon integrated circuits, the concepts are equally applicable to other types of devices susceptible to electrical stress effects. For example, such transient suppressor devices find utility in large-scale integrated circuits and microprocessors, in electro-optical devices, and in microwave acoustic and micromechanical devices, and in other applications.
• The disclosed MOS transient suppression device exemplifies advantages and characteristics fulfilling long apparent needs within the industry. These benefits include, among other things, programmability within appropriate ranges via simple adjustments during microfabrication, relatively high switching speed, relatively high peak current-carrying capability, bidirectionally symmetric performance, and ready compatibility with standard CMOS processing parameters, equipment and design rules.
• Although the description above employs language specific to exemplary structural features and/or methodological acts, it is to be understood that the appended claims are not limited to the specific features or acts described. Rather, the specific features and acts are disclosed as exemplary forms of implementing these disclosed concepts.

Claims (4)

1-140. (canceled)

141. A voltage stress protection circuit, comprising:

a semiconductive substrate; and

a clamp circuit formed on the substrate and having first and second conductors, the clamp circuit being configured to provide a clamp voltage of about seven volts or less with respect to a voltage difference between the first and second conductors, the clamp circuit including clamping characteristics that are symmetric about a predetermined voltage, wherein the clamp circuit comprises a five-layer NPNPN structure.

142. The voltage stress protection circuit ofclaim 141, wherein the clamp circuit comprises five layers of doped semiconductive material.

143. The voltage stress protection circuit ofclaim 141, wherein the first conductor of the clamp circuit is capacitively isolated from the substrate via a structure having a dielectric constant of 3.9 or less.

Images