BIPOLAR TRANSISTOR HAVING SINKER DIFFUSION UNDER A TRENCH
[0001] This relates in general to integrated circuits, and in particular to a bipolar transistor having sinker diffusion under a trench.
BACKGROUND
[0002] Bipolar junction transistors are active semiconductor devices formed by a pair of P-N junctions, including an emitter-base junction and a collector-base junction. An NPN bipolar junction transistor has a thin region of P-type material providing the base region between two regions of N-type material that provide the emitter and collector regions. A PNP bipolar junction transistor has a thin region of N-type material providing the base region between two regions of P-type material that provide the emitter and collector regions. The movement of electrical charge carriers, which produces electrical current flow between the collector and emitter regions, is controlled by an applied voltage across the emitter-base junction.
[0003] For electrostatic discharge (ESD) protection of other devices, conventional vertical NPN bipolar devices typically include an n-buried layer (NBL) together with a N+ sinker diffusion as collectors in each device stripe, which provides a low resistance-path to conduct ESD strike-induced current back to the top surface of the substrate (e.g., a top silicon surface). When forming the N+ sinker, an ion implant is followed by a thermal diffusion of the dopant to reach a depth of several micrometers for providing a low resistance connection to the NBL. Due to a large lateral diffusion of the N+ sinker, the device pitch is limited to a large dimension, which consumes significant area on the integrated circuit (IC) die.
SUMMARY
[0004] In described examples, a bipolar transistor includes: a substrate having a semiconductor surface; and first and second trench enclosures, both at least lined with a dielectric extending downward from a topside of the semiconductor surface to a trench depth. The first trench enclosure defines an inner enclosed area. A base and an emitter formed in the base are within the inner enclosed area. A buried layer is below the trench depth, including under the base. A sinker diffusion includes a first portion between the first and second trench enclosures, extending from the topside of the semiconductor surface to the buried layer, and a
second portion within the inner enclosed area. The second portion does not extend to the topside of the semiconductor surface.
BRIEF DESCRIPTION OF THE DRAWINGS
[0005] FIG. 1 is a cross-sectional view of an example vertical bipolar transistor having a sinker diffusion that up-diffuses around first and second trench enclosures, which confines the lateral diffusion of the sinker to build the collector of the vertical bipolar transistor, according to an example embodiment.
[0006] FIG. 2 is a cross-sectional view of an example vertical bipolar transistor having a sinker diffusion that up-diffuses around first and second trench enclosures, which confines the lateral diffusion of the sinker to build the collector of the vertical bipolar transistor, along with a third trench enclosure that forces the breakdown diode triggered by an ESD event to be laterally shifted away (and thus separated) from the vertical bipolar transistor's active region, according to an example embodiment.
[0007] FIG. 3 is a block diagram of an ESD protected integrated circuit (IC), into which disclosed vertical bipolar transistors may be incorporated to protect one or more terminals of the IC, according to an example embodiment.
[0010] FIG. 4A is a diagram of current flow from emitter (E) to collector (C) for an NPN vertical bipolar transistor having a conventional trench-less N+ sinker diffusion as the collector, upon receiving a simulated ESD strike, with the lines shown depicting the ESD induced current flow.
[0011] FIG. 4B is a diagram of current flow from emitter (E) to collector (C) for the NPN vertical bipolar transistor of FIG. 2, which confines the lateral diffusion of the sinker to provide the collector of the vertical bipolar transistor, along with a third trench enclosure.
DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS
[0008] In described examples, vertical bipolar transistors include a sinker diffusion that up-diffuses around a pair of first and second trench enclosures, which confines the lateral diffusion of the sinker to provide the collector of the vertical bipolar transistor. This disclosed confining of the lateral diffusion of the sinker significantly reduces the lateral size of the sinker diffusion and thus the footprint of the transistor.
[0009] Another trench enclosure may be added outside the pair of trench enclosures. In that manner, for electrostatic discharge (ESD) protection applications, the breakdown diode that is
triggered by an ESD event is laterally shifted away and thus separated from the vertical bipolar transistor's active region. The addition of the extra trench enclosure enables better control of the ESD induced current, and it forces the resulting hot spot deeper from the top surface of the substrate. This increases robustness under extreme ESD events, such as IEC (e.g., IEC 61000-4-2 standard) pulses.
[0012] FIG. 1 is a cross-sectional view of an example NPN vertical bipolar transistor 100 having a N+ sinker diffusion 115 that up-diffuses around a first trench enclosure 121 and a second trench enclosure 122, which confines the lateral diffusion of the N+ sinker diffusion 115 to build the collector of the vertical bipolar transistor, according to an example embodiment. Although the disclosed vertical bipolar transistors are generally described as NPNs, PNP bipolar transistors can be formed by reversing the doping types along with other minor design modifications.
[0013] Moreover, although generally shown for simplicity as a single bipolar device, an array of the disclosed bipolar transistors is useful for practical ESD protection applications, as compared to a single large area bipolar transistor. For example, the disclosed collectors lead to a perimeter device, because the collector is formed by up-diffusion around a pair of trench enclosures. Accordingly, scaling of a single NPN bipolar unit cell generally has limited efficacy, as eventually the resistance of the N+ sinker diffusion 115 or the n-buried layer (NBL) 126 will limit the current handling capability. An array of unit bipolar cells will generally be advantageous for current handling capability, but with some penalty in area consumed by the N+ sinker diffusion and the trench enclosures.
[0014] First trench enclosure 121 and second trench enclosure 122 are at least dielectric lined and extend down from the topside 106a of the semiconductor surface 106 of the substrate shown as a P- substrate 105. The substrate 105 and/or semiconductor surface 106 can include silicon, silicon-germanium (SiGe), or other semiconductor material. One particular arrangement is a SiGe semiconductor surface on a silicon substrate. For example, first trench enclosure 121 and second trench enclosure 122 may include shallow trench isolation (STI), or other suitable dielectric isolation structures. The first trench enclosure 121 defines an inner enclosed area. The trench depths for first trench enclosure 121 and second trench enclosure 122 are both typically in the range from 0.5 μιη to 8 μιη.
[0015] The first and second trench enclosures 121 and 122 can be filled with dielectric (such
as silicon oxide, silicon nitride or silicon oxynitride), or lined by a dielectric and then filled with another material (such as polysilicon). An example trench process includes forming a pattern and then etching a hole in a hard mask (such as a pad oxide, plus silicon nitride or thicker layers if needed), dry etching to cut the trenches, and a thermal liner oxidation to form a first dielectric layer. Deposited oxide can further fill the trenches, and deposited polysilicon (doped or undoped) can also be used in filling the dielectric lined trenches.
[0016] Vertical bipolar transistor 100 includes a base 140 and an emitter 150 formed in the base. Although a single emitter 150 is shown, the disclosed bipolar transistors can have multiple emitters. Emitter 150 can include phosphorous, arsenic or antimony. An NBL 126 is below the trench depth, including under the base 140. NBL 126 generally includes phosphorous, but may also include other n-dopants. A contact 158 is shown to the P+ region 148 for contacting the base 140. A contact 154 is shown to P+ region 146 for contacting the semiconductor surface 106 and substrate 105. A contact 156 is shown to emitter 150. A contact 160 is also shown to the topside surface of the N+ sinker diffusion 115. A dielectric layer 167 is lateral to the respective contacts on the topside 106a of the semiconductor surface 106.
[0017] The N+ sinker diffusion 115 includes a first portion 115a between the first and second trench enclosures 121 and 122, extending from the topside 106a of the semiconductor surface 106 to the NBL 126, and a second portion 115b within the enclosed area defined by the first trench enclosure 121. The second portion 115b does not extend to the topside 106a of the semiconductor surface 106. The pair of trench enclosures 121 and 122 confine the N+ sinker diffusion 115, causing it to diffuse in a one-dimensional (ID) manner, which allows it to become deeper for the same thermal cycle as compared to diffusing it in 2 dimensions (2D).
[0018] The transistor 100 uses current flow through the N+ sinker diffusion 115, which is between trench enclosures 121 and 122, to conduct current back to the topside 106a of semiconductor surface 106. As a result of this vertical architecture, a "hot spot" at the base-collector junction is deep within the semiconductor surface 106 or the substrate 105 (e.g., silicon), resulting in a full 4π steradians of solid angle to diffuse away the heat, resulting in good robustness to high current pulses when used as an ESD protection device. Moreover, because the N+ sinker diffusion 115 is between the trench enclosures 121 and 122, the lateral diffusion of the sinker is reduced, and the resulting transistor footprint is compact.
[0019] In contrast, conventional NPNs for ESD protection are built using a conventional
trench-less N+ sinker diffusion to NBL connection as the buried collector. Such design can be problematic, because: (a) the N+ sinker diffuses laterally for several micrometers; and (b) a significant portion of the ESD induced current flows near the surface, which causes a localized failure at low ESD power densities. These two factors combine to make conventional NPNs for ESD protection very large in area for a given ESD rating.
[0020] FIG. 2 is a cross-sectional view of an example vertical bipolar transistor 200 having a disclosed N+ sinker diffusion 115 that up-diffuses around a first and a second trench enclosure 121 and 122, which confines the lateral diffusion of the sinker diffusion 115 to build the collector of the vertical bipolar transistor, along with a third trench enclosure 123 that forces the breakdown diode triggered by an ESD event to be laterally shifted away (and thus separated) from the vertical bipolar transistor's active region, according to an example embodiment. The base shown as 140' in this embodiment extends outside of the third trench enclosure 123. In this embodiment, the breakdown is set by the lateral avalanche diode having a junction at the edge of the base 140' between the first trench enclosure 121 and third trench enclosure 123, which in some embodiments is 15 V to 20 V.
[0021] The ability to survive an ESD event is a key specification for integrated circuits (ICs). A conventional method for providing such ESD protection is to include one or more ESD clamping devices that are connected across the external pins of an IC. More generally, the ESD protection or clamping devices are connected between the input terminals of, and thus in parallel with, the circuitry to be protected. These clamping devices are generally designed to breakdown at a voltage below that which would cause damage to the internal circuitry of the IC, thus absorbing the ESD energy and protecting the IC circuitry. The disclosed vertical bipolar transistors can be designed to operate in the bipolar snapback mode to protect the IC circuitry, such as metal oxide semiconductor (MOS) transistors on the IC.
[0022] The disclosed vertical bipolar transistors are suitable for a variety of applications including ESD protection, such as for protecting power MOS transistors or power bipolar transistors. In contrast, a conventional solution for ESD protection of MOS transistors is to use an NPN vertical bipolar transistor having a trench-less N+ sinker diffusion as the collector sinker. Such a device is large in area, due to the N+ sinker's lateral diffusions. The robustness of such conventional devices for ESD protection is also poor, due to the shallow depth of the hot spot at failure (as described below in connection with FIG. 4A). Another conventional solution
is to use a lateral NPN bipolar, but in this case the hot spot at failure is near the top surface, so the robustness is also poor. As described above, in contrast, the disclosed confining of the lateral diffusion of the sinker by a pair of trench enclosures significantly reduces the lateral size of a sinker diffusion, which enables a compact layout, a deeper hotspot, and improved robustness during ESD strikes (as described below in connection with FIG. 4B).
[0023] The base doping may be designed with a breakdown voltage for providing ESD protection to target devices. For example, base doping may be added (increased) for the disclosed bases, such as base 140 for vertical bipolar transistor 100 in FIG. 1 and base 140' for vertical bipolar transistor 200 in FIG. 2, to lower the breakdown voltage of the device. In one embodiment, the disclosed bipolar transistors are co-fabricated with MOS transistors, such as with one or more laterally diffused power MOS transistors (e.g., laterally diffused MOS (LDMOS) or double-diffused MOS (DMOS)). In this embodiment, a well (DWELL) diffusion (including boron for the body of the LDMOS or DMOS devices) can also be used for the base of the disclosed vertical bipolar transistors. Adding such a DWELL diffusion to a disclosed vertical bipolar transistor is a convenient way to increase the doping level of the more lightly doped base side of the avalanche junction, reducing the breakdown voltage to ~7V or 8V or less, to be more suitable for protection of a 5V device.
[0024] In the LDMOS or DMOS process flow, an SNWELL or DNWELL (a deeper, less heavily doped diffusion) can be used around the drain of an NMOS transistor to increase the operating voltage. Similarly, an SPWELL can be used around the drain of a PMOS transistor. An SPWELL also can be used as the body of an NMOS transistor (DWELL is in the body of the LDMOS, so SPWELL would not be used there). SNWELL can be used as the body of a PMOS.
[0025] The breakdown diode in a disclosed 20V (or 15V or 25V) NPN bipolar transistor can be set by a lateral SNWELL-SPWELL avalanche diode. The separation between the photomask edges for delimiting the SNWELL and SPWELL implants can be abutted (zero SNWELL-SPWELL separation) to reduce the breakdown voltage, or separated by up to 1 to 2 micrometers to increase the breakdown voltage. Abutted SNWELL-SPWELL junctions produce an avalanche breakdown voltage of ~11V to 15V. The breakdown can be increased by several volts by increasing the SNWELL-SPWELL separation, such as by 1 or 2 micrometers.
[0026] A 5V NPN requires a lower breakdown junction, such as lower than provided by conventional NSD-SPWELL or PSD-SNWELL junctions. The breakdown of these junctions is
typically in the range of 7V to 11V. As stated above, this range is wide and contains values far in excess of 5V, so using a DWELL to increase the base doping lowers the breakdown voltage by up to a few volts, which can be useful for ESD protection of 5V MOS devices.
[0027] FIG. 3 illustrates a high level depiction of an ESD protected IC into which vertical bipolar transistor 100 may be incorporated for protecting one or more terminals of the IC, according to an example embodiment. The "T" indicated at the top of the respective ESD protection devices 100 in FIG. 3 represents an input from a suitable trigger circuit.
[0028] IC 300 is shown including functional circuitry 324, which can include complementary metal-oxide-semiconductor (CMOS) functional circuitry and/or bipolar functional circuitry, which is circuitry that realizes and performs desired functionality of IC 300, such as that of a digital IC (e.g., digital signal processor) or analog IC (e.g., amplifier or power converter). The capability of functional circuitry 324 provided by IC 300 may vary, ranging from a simple device to a complex device.
[0029] IC 300 also includes a number of external terminals, through which functional circuitry 324 performs its operations. A few of those external terminals are illustrated in FIG. 3. The number of terminals and their function can also vary widely. In the example of IC 300, two terminals operate as common input and output terminals (I/O), through which functional circuitry 324 can receive incoming signals and can generate outputs. FIG. 3 also shows a dedicated input terminal IN and a dedicated output terminal OUT. Each of terminals IN and OUT are also connected to functional circuitry 324. Power supply terminal Vdd receives a positive power supply voltage in this example, while ground terminal Vss is provided to receive a reference voltage, such as system ground. Although not shown, the ground shown connected to the ESD protection devices 100 is coupled to Vss, such as by resistively coupling or shorting together.
[0030] IC 300 includes a disclosed vertical bipolar transistor 100 connected to each of its terminals. Each vertical bipolar transistor 100 is connected to its corresponding terminal in parallel with the functional circuitry 324. Vertical bipolar transistors 100 are also connected to power supply and reference voltage terminals VDD and Vss in parallel with functional circuitry 324. However, in some applications, some pins of the protected device will be self-protecting, such as diode protected power supply pins. Pins also can be protected against different levels of ESD strike (such as human body model (HBM), charged device model (CDM), and IEC).
[0031] FIG. 4A depicts the current flow from emitter (E) to collector (C) for an NPN vertical bipolar transistor having a conventional trench-less N+ sinker diffusion as the collector, upon receiving a simulated ESD strike, with the lines shown depicting the ESD induced current flow. The lines of current near the top surface of the transistor evidence a lack of robustness upon an ESD strike, due to the shallow depth of the resulting ESD induced hot spot failure.
[0032] FIG. 4B depicts the current flow from emitter (E) to collector (C) for the NPN vertical bipolar transistor 200 of FIG. 2 having a disclosed sinker diffusion that up-diffuses around a first and a second trench enclosure, which confines the lateral diffusion of the sinker to provide the collector of the vertical bipolar transistor 200, along with a third trench enclosure. The lines of current shown evidence no lines of ESD induced current near the top surface, with the lines of current shown evidencing the breakdown diode triggered by the simulated ESD event also being laterally shifted away (and thus separated) from the vertical bipolar transistor's active region. As described above, in this embodiment, the device breakdown is set by the lateral avalanche diode having a junction at the edge of the base (140' in FIG. 2) between the first trench enclosure (shown in FIG. 4B as 12 Γ to represent a polysilicon- filled version of first trench enclosure 121 of FIG. 2) and third trench enclosure (shown as 123' to represent a polysilicon-filled version of the third trench enclosure 123 of FIG. 2).
[0033] The disclosed embodiments can be used to form semiconductor die that may integrated into a variety of assembly flows, in order to form a variety of different devices and related products. The semiconductor die may include various elements therein and/or layers thereon, including barrier layers, dielectric layers, device structures, active elements and passive elements, such as source regions, drain regions, bit lines, bases, emitters, collectors, conductive lines, and conductive vias. Moreover, the semiconductor die can be formed from a variety of processes, such as bipolar, CMOS, BiCMOS and MEMS.
[0034] Modifications are possible in the described embodiments, and other embodiments are possible, within the scope of the claims.