Switch circuit with controllable phase node ringing

Abstract

A switch circuit includes a first MOS transistor and a second MOS transistor of a same conductivity type connected in parallel between a first terminal and a second terminal of the switch circuit, the first and second MOS transistors having respective gate terminals coupled to the control terminal to receive a control signal to turn the first and second MOS transistors on or off. The first MOS transistor is characterized by a first reverse gate-to-drain capacitance (Crss) and the second MOS transistor is characterized by a second Crss that is greater than the first Crss.

Description

FIELD OF THE DISCLOSURE

This disclosure relates to integrated circuits and more specifically to integrated circuit devices having metal-oxide-semiconductor field effect transistors (MOSFETs).

BACKGROUND OF INVENTION

Integrated circuits, such as microprocessors and memory devices, contain many MOSFET transistors on each device, providing the basic switching functions required to implement logic gates and/or data storage. In one example, a DC-to-DC power converter, such as a buck converter, usually employs MOSFETs as its switching devices. In particular, a buck converter widely used throughout the industry converts a higher input voltage into a lower output voltage. A synchronous buck converter includes a pair of switching devices coupled in series across the input voltage source. The switching devices are typically MOSFETs. FIG. 1 shows a conventional buck converter which consists a high-side and a low-side switch. The high-side switch 102 is coupled to the voltage supply (Vin) and the low-side switch 104 is connected to ground. An output filter typically including an inductor (L) and a capacitor (Cout) is connected to a junction (i.e., phase node or switch node) 105 formed by the pair of switches 102 and 104 for providing the output voltage to a load. A controller 110 drives the switches to connect the output filter to the voltage supply or to ground to maintain the output voltage at a predetermined level.

One issue in a synchronous buck converter design is voltage overshoots and ringing on the phase node. Due to the growing demand for higher performance devices, the high side MOSFET's switching speed are expected to be faster. The faster the switching speed, the faster the voltage transition of the phase node. The fast voltage transition may cause the inadvertent turn-on of the low-side switch, resulting in a converter malfunction and reducing efficient. In addition, it may cause the phase node ringing as the magnitude of the ringing is a function of the high side MOSFET's switching speed. The phase node voltage spikes can induce malfunction or undesired stress of the power switches.

A number of techniques are used to control phase node ringing. As shown in FIG. 1, the boost resistor RBOOT and the high side gate resistor RGH have been included in the synchronous buck converter of FIG. 1 for control of the phase node ringing phenomenon. Specifically, the boost resistor RBOOT can slow down the turn-on of the high side MOSFET without affecting the turn-off. In addition, the high side gate resistor RGH in series with the gate can increase both the turn-on and turn-off times of the high-side MOSFET, which controls ringing on the rise and fall of the phase node. Another proposal is to integrate a Schottky diode in the low-side switch to reduce recovery charge, thus avoiding voltage spikes at phase node. It is desirable to design a switch circuit that can properly control the phase node ringing of a synchronous buck converter.

It is within this context that embodiments of the present invention arise.

BRIEF DESCRIPTION OF THE DRAWINGS

Objects and advantages of aspects of the present disclosure will become apparent upon reading the following detailed description and upon reference to the accompanying drawings in which:

FIG. 1 is a schematic diagram of a conventional buck converter.

FIG. 2 is a circuit diagram of a switch circuit according to aspects of the present disclosure.

FIG. 3 is a cross-sectional schematic diagram of portions of a conventional trench MOSFET.

FIG. 3A is an example layout of a switch circuit of FIG. 2.

FIGS. 4A-4J are cross-sectional views illustrating the fabrication process for forming a switch circuit of FIG. 2 in one embodiment of present disclosure.

FIGS. 4K-4N are cross-sectional views illustrating the fabrication process for forming a switch circuit of FIG. 2 in one embodiment of present disclosure.

FIGS. 5A-5J are cross-sectional views illustrating the fabrication process for forming a switch circuit of FIG. 2 in an alternative embodiment of present disclosure.

FIG. 6 is a circuit diagram of a switch circuit according to aspects of the present disclosure.

FIG. 7 is a waveform showing the phase node waveform of a switch circuit of FIG. 6.

FIG. 8 is a schematic diagram of a buck converter in accordance with aspects of the present disclosure.

DESCRIPTION OF THE SPECIFIC EMBODIMENTS

Aspects of the present disclosure propose to replace the conventional low-side MOSFET switch with a switch circuit that allows a synchronous buck converter to have a smooth phase node waveform. Aspects of the present disclosure include a switch circuit having a main MOS transistor and a secondary MOS transistor connected in parallel. The main MOS transistor is characterized by a first reverse gate-to-drain capacitance (C_rss) and a first threshold voltage. The secondary MOS transistor is characterized by a second C_rss and a second threshold voltage. In some embodiments, the phase node voltage ringing of a synchronous buck converter using such a switch circuit can be reduced by increasing the C_rss of the secondary MOS transistor such that the second C_rss is greater than the first C_rss. In some embodiments, the phase node voltage ringing can be further reduced by lowering the threshold voltage of the secondary MOS transistor such that the second threshold voltage is lower than the first threshold voltage. In some other embodiments, the phase node voltage ringing can be reduced by increasing the C_rss of the secondary MOS transistor and lowing the threshold voltage of the secondary MOS transistor such that the second C_rss is greater than the first C_rss and the second threshold voltage is lower than the first threshold voltage.

Reverse Source-to-Source Capacitance (C_rss)

One approach to smooth the phase node waveform of a synchronous buck converter is to replace the conventional single low-side MOSFET with a main MOS transistor and a parallel secondary MOS transistor having a higher C_rss comparing to that of the main MOS transistor. FIG. 2 illustrates a switch circuit 200 that may be used as, e.g., the low-side switch 104 in the buck converter shown FIG. 1, according to aspects of the present disclosure. The switch circuit 200 has a first terminal, e.g., a drain, D, a second terminal, e.g., a source, S and a control terminal, e.g., a gate, G. The switch circuit 200 has two MOSFETs, a main MOS transistor 210 having regular C_rss and a secondary MOS transistor 220 having high C_rss connected in parallel between the first terminal D and the second terminal S. The main and the secondary MOS transistors 210 and 220 have respective gate terminals coupled to the control terminal G to receive a control signal to turn the main and the secondary MOS transistors 210, 220 on or off.

The main MOS transistor 210 and the secondary MOS transistor 220 are of the same conductivity. The main MOS transistor 210 and the secondary MOS transistor 220 are formed in a common active device area of a common semiconductor substrate. Specifically, each of the main MOS transistor 210 and the secondary MOS transistor 220 comprises one or more trench transistors. Each of the trench transistors has a gate terminal formed in a trench in a semiconductor layer on the common semiconductor substrate, a body region formed in the semiconductor layer and a source region formed proximate the trench.

An area of the main MOS transistor 210 is a regular C_rss area, and an area of the secondary MOS transistor 220 is a high C_rss area. The ratio of the high C_rss area to the regular Cr_ss area affects the phase node peak voltage. The higher area ratio results in a lower phase node peak voltage, and thus smoothing the phase node waveform. Alternatively, the higher the C_rss in the high C_rss area, the smaller the area ratio needed to get the desired effect of reduced phase node ringing. The ratio may be pre-determined according to the desired reduction on phase node peak voltage. By way of example and not by way of limitation, the high C_rss area may be between 0.1% and 15% of a total area of the common active device area.

FIG. 3 shows a cross-sectional schematic diagram of portions of a conventional trench MOSFET. As well known in the industry, C_rss is basically proportional to the depth between the bottom of the body region 320 and the bottom of the gate trench 340 as indicated in FIG. 3. Thus, one approach to increase C_rss is to increase the trench depth for the gate trenches 340 in the high Crss area. Another approach is to lower the depth of the body region 320 in the high C_rss area of the active region. It should be noted that, in some embodiments, the C_rss can be increased by having a deeper gate trenches in conjunction with a shallower body region in the high C_rss area.

Preferably the area of the secondary MOS transistor 220 should be evenly distributed through the area of the main MOS transistor 210. FIG. 3A is an example layout of the switch circuit 200 formed on a semiconductor substrate including the main MOS transistor 210 and the secondary MOS transistor 220 without showing a top source metal layer. The gate trenches are in parallel stripes with source regions disposed on top of the substrate between the parallel gate trenches with contact openings opened through the source regions. Segments of the secondary MOS transistor 220 are disposed along each stripe MOS structure forming stripes of secondary MOS transistor 220 in a direction perpendicular to the gate trench stripe direction. Alternatively, an entire length of certain gate trenches may be selected to form a stripe of the secondary MOS transistor 220 in the direction of the gate trench stripe.

First Embodiment

FIGS. 4A-4J are cross-sectional views illustrating the fabrication process for forming a switch circuit of FIG. 2 having the gate trench for each trench transistor of the secondary MOS transistor 220 deeper in depth in the semiconductor layer than the gate trenches of the trench transistors of the main MOS transistor 210.

Referring to FIG. 4A, the process uses a semiconductor substrate 410 of a first conductivity as a starting material. In some embodiments, the substrate 410 may include an N-type epitaxial layer on top of a heavily doped N type (N+) silicon wafer. A mask 412 applied on the substrate 410 includes openings to define locations of a plurality of gate trenches for the trench transistors of the main MOS transistor 210 and the secondary MOS transistor 220. In FIG. 4B, an etching process is performed and the corresponding portions of the underlying substrate 410 are etched down to form a plurality of the gate trenches 440.

Another mask 414 is applied to mask the gate trenches 440a in the regular Crss area (i.e., the gate trenches for the trench transistors of the main MOS transistor 210) and uncover the gate trenches 440b in the high Crss area (i.e., the gate trenches for the trench transistors of the secondary MOS transistor 220). Another etch process is performed to make the gate trenches 440b deeper as shown in FIG. 4C. The mask 414 may then be removed, as shown in FIG. 4C′. Subsequent processing may proceed as described below with respect to FIGS. 4D-4J.

In an alternative implementation, shown in FIGS. 4K-4N, trenches of two different depths may be formed by modifying a mask between two etch processes. Specifically, as shown in FIG. 4K, a mask 412 having a first set of openings 413 is applied on the substrate 410. A first etching process is performed, as shown in FIG. 4L to etch gate trenches 440b into the substrate to an initial depth. Additional openings 413′ are then formed in the mask 412, as shown in FIG. 4M. Alternatively, the mask 412 may be removed and a second mask may be formed having openings corresponding to the original openings 413 and the additional openings 413′. A second etch process etches the substrate 410 through the additional openings 413′ to form gate trenches 440a while the second etch process also etches the substrate through the original openings 413 to deepen the trenches 440b. The mask 412 may then be removed and subsequent processing may proceed as described below with respect to FIGS. 4D-4J.

Specifically, after the trenches 440a, 440b have been formed to their respective depths and the masks have been removed, a sacrificial oxide layer (not shown) may be grown and then removed to improve the silicon surface. Referring to FIG. 4D, an insulating layer (e.g., gate oxide) 444 is then formed along the inner surface of the gate trenches 440a and 440b. A conductive material is then deposited over the gate oxide layer 444 with an etch back to form the gate electrode 442. In some embodiments, the conductive material can be in-situ doped or undoped polysilicon. A dielectric layer 446 may be formed above the trenches to cap the trenches and above the substrate 410 as an implant buffer layer.

Referring to FIG. 4E, a blanket body implant is performed to form body regions 420. The dopant ions are of the opposite conductivity type to the doping of the substrate 410. In some embodiments, the dopant ions can be Boron ions for an N-channel device. In some embodiments, Phosphorous or Arsenic ions can be used for P-channel devices. Heat is then applied to activate dopant atoms and drive dopant diffusion to form the body regions 420 as shown in FIG. 4F.

Referring to FIG. 4G, source implant and diffusion are performed to form source regions 430 in the top surface of the substrate 410. The dopant ions are of the same conductivity type to the doping of the substrate 410. In some embodiments, Arsenic ions can be implanted for an N-channel device. Alternatively, Boron ion can be implanted for a P-channel device.

Subsequently, a planarized dielectric layer 450 is deposited over the substrate 410 as shown in FIG. 4H. In some embodiments, the dielectric layer 450 is formed by a low temperature oxidation process with material of Borophosphorosilicate Glass (BPSG).

A contact photoresist (not shown) is then applied on the dielectric layer 450 with a pattern that has an opening at the locations of contact trenches. An etch process is performed to remove the uncovered portions of the dielectric layer 450 and form contact trenches 460 through the source region 430 into the body region 420 as shown in FIG. 4I

A metal layer 470 is then deposited to fill the contact openings and to interconnect all the source regions to form the main MOS transistor 210 and secondary MOST transistor 220 connected in parallel.

Second Embodiment

FIGS. 5A-5J are cross-sectional views illustrating the fabrication process for forming a switch circuit of FIG. 2 by making the body region for each trench transistor of the secondary MOS transistors 220 shallower in depth comparing to the body regions for the trench transistors of the main MOS transistors 210.

Referring to FIG. 5A, the process uses a semiconductor substrate 510 of a first conductivity as a starting material. In some embodiments, the substrate 510 may include an N-type epitaxial layer on top of a heavily doped N type (N+) silicon wafer. A mask 512 applied on the substrate 510 includes openings to define locations of the gate trenches for the trench transistors of the main MOS transistor 210 and the secondary MOS transistor 220. As shown in FIG. 5B, an etching process is performed and the corresponding portions of the underlying substrate 510 are etched down to form a plurality of the gate trenches 540. The mask 512 may then be removed, as shown in FIG. 5B′.

Next, a sacrificial oxide layer (not shown) may be grown and then removed to improve the silicon surface. An insulating layer (e.g., gate oxide) 544 is then formed along the inner surface of the gate trenches 540. A conductive material is then deposited over the gate oxide layer 544 with an etch back to form the gate electrode 542. In some embodiments, the conductive material can be in-situ doped or undoped polysilicon. A dielectric layer 546 may be formed above the trenches to cap the trenches and above the substrate 510 as an implant buffer layer as shown in FIG. 5C.

Referring to FIG. 5D, a first blanket body implant with a lower energy is performed to form body regions 520. The dopant ions are of the opposite conductivity type to the doping of the substrate 510. In some embodiments, the dopant ions can be Boron ions for an N-channel device. In some embodiments, Phosphorous or Arsenic ions can be used for P-channel devices. In some embodiments, the dopant ions are implanted into the body regions at a dose of, for example, Boron ions may be implanted to a dose of, e.g., about 1e13 cm⁻² and an energy of, for example about 100 KeV. After the first low-energy body implant, selected body regions 520a corresponding to the high C_rss area (i.e., the body regions for the trench transistors of the secondary MOS transistor 220) are protected by an implant mask 522. Another body implant at higher energy is performed in FIG. 5E though openings in the mask. In other words, the dopant ions are only implanted into the body regions 520b in the regular C_rss area (i.e., the body regions for the trench transistors of the main MOS transistor 210) in the second body implant. As a result, the body region 520b is deeper in depth into the substrate 510 comparing to the body region 520a in the high C_rss area. In some embodiments, the second body implant the dopant (e.g., Boron) ions into the body regions 520b at a dose of, for example, about 8e12 cm⁻² and an energy of, for example about 150 KeV. Heat is then applied to activate dopant atoms and drive dopant diffusion to form the body regions 520a and 520b as shown in FIG. 5F.

Referring to FIG. 5G, source implant is performed to form source regions 530 in the top surface of the substrate 510. The dopant ions are of the same conductivity type to the doping of the substrate 510. In some embodiments, Arsenic ions can be implanted for an N-channel device. Alternatively, Boron ion can be implanted for a P-channel device.

Subsequently, a planarized dielectric layer 550 is deposited over the substrate 510 as shown in FIG. 5H. In some embodiments, the dielectric layer 550 is formed by a low temperature oxidation process with material of Borophosphorosilicate Glass (BPSG).

A contact photoresist (not shown) is then applied on the dielectric layer 550 with a pattern that has an opening at the locations of contact trenches. An etch process is performed to remove the uncovered portions of the dielectric layer 550 and form contact trenches 560 through the source region 530 into the body region 520a and 520b as shown in FIG. 5I

A metal layer 570 is then deposited to fill the contact openings and to interconnect all the source regions to form the main MOS transistor 210 and MOS transistor 220 connected in parallel.

Threshold Voltage

Another approach smoothes the phase node waveform of a synchronous buck converter by replacing the conventional single MOSFET with a main MOS transistor and a parallel secondary MOS transistor with a lower turn-on threshold voltage (Vth). This approach can keep the phase node peak ringing low because the secondary MOS transistor turns on due to its turn-on threshold voltage lower than the phase node voltage spike induced gate spike. Specifically, when the high-side MOSFET turns on, the current drops in the low-side FET. A spike in VDS triggers a high enough VGS spike, and it turns on the secondary MOS transistor with a lower-turn-on threshold voltage and thus reducing the voltage of the phase node spikes. FIG. 6 illustrates a switch 600 circuit according to aspects of the present disclosure. The switch circuit 600 having a first terminal D, a second terminal S and a control terminal G. The switch circuit 600 has two MOSFETs, a main MOS transistor 610 having a regular Vth and a secondary MOS transistor 620 having a lower Vth connected in parallel between the first terminal D and the second terminal S. The main and the secondary MOS transistors 610 and 620 have respective gate terminals coupled to the control terminal G to receive a control signal to turn the main and the secondary MOS transistors 610, 620 on or off.

The main MOS transistor 610 and the secondary MOS transistor 620 are of the same conductivity. The main MOS transistor 610 and the secondary MOS transistor 620 are formed in a common active device area of a common semiconductor substrate. Specifically, each of the main MOS transistor 610 and the secondary MOS transistor 620 comprises one or more trench transistors. Each of the trench transistors has a gate terminal formed in a trench in a semiconductor layer on the common semiconductor substrate, a body region formed in the semiconductor layer and a source region formed proximate the trench. An area of the main MOS transistor 610 is a regular Vth area, and an area of the secondary MOS transistor 620 is a lower Vth area. The ratio of the lower Vth area to the regular Vth area affects the phase node peak voltage. FIG. 7 is a waveform showing that the phase node peak ringing becomes lower with more percentage of the low Vth area in a total area of a common active device area. The ratio may be pre-determined according to the desired reduction on phase node peak voltage. By way of example and not by way of limitation, the low Vth area is between 5% and 15% of a total area of the common active device area. The lower the Vth of the secondary MOS transistor, the lower the area ratio needed to get the desired effect of reduced phase node ringing.

As to the fabrication process for a switch circuit in FIG. 6, two body implants are performed. Specifically, a regular blanket body implant is first performed followed by a second body implant using a body mask. The body mask covers the low Vth area so that dopants ions are only implanted into the regular Vth area in the second body implant. It is noted that threshold voltage can be reduced by low implant energy and can be further reduced by reducing the dose for implant. In some embodiments, the dopant ions (e.g., Boron ions) are implanted into the body regions at a dose of, for example, from about 2e12 cm⁻² to about 2e13 cm⁻² and an energy of, for example about 50 KeV 1 MeV in the first body implant, and at a dose of, for example, about 2e12 cm⁻² to about 2e13 cm⁻² and an energy of, for example about 80 KeV-1 MeV in the second body implant. Details of the fabrication process can be found in commonly owned U.S. patent application Ser. No. 14/252,568 filed Apr. 14, 2014, published as U.S. Patent Application Publication 2015/0295495, which describes a switch circuit having a parallel-connected transistor pair; one has a lower threshold voltage than the other. The entire disclosures of U.S. Patent Application Publication 2015/0295495 are incorporated herein by reference. As depicted in FIG. 8, switch circuits of the types described herein may be used in buck converters, e.g., as the low-side switch 104 shown in FIG. 1. The buck converter may include e.g., the high-side switch 102 with its own input (e.g., drain) terminal D, output (e.g., source) terminal S, and a low-side control terminal (e.g., gate) G. The input of the low-side switch is connected to the second (output) terminal of the high side switch.

While the above is a complete description of the preferred embodiments of the present invention, it is possible to use various alternatives, modifications, and equivalents. Therefore, the scope of the present invention should be determined not with reference to the above description but should, instead be determined with reference to the appended claims, along with their full scope of equivalents. Any feature, whether preferred or not, may be combined with any other feature, whether preferred or not. In the claims that follow, the indefinite article “A”, or “An” refers to a quantity of one or more of the item following the article, except where expressly stated otherwise. The appended claims are not to be interpreted as including means-plus-function limitations, unless such a limitation is explicitly recited in a given claim using the phrase “means for.” Any element in a claim that does not explicitly state “means for” performing a specified function, is not to be interpreted as a “means” or “step” clause as specified in 35 USC § 112, ¶6

Claims

1. A switch circuit having a first terminal, a second terminal and a control terminal, the switch circuit comprising: a first MOS transistor and a second MOS transistor of a same conductivity type connected in parallel between the first terminal and the second terminal, the first and second MOS transistors having respective gate terminals coupled to the control terminal to receive a control signal to turn the first and second MOS transistors on or off, wherein the first and second transistors are formed in a common active device area of a common semiconductor substrate,wherein the first MOS transistor is characterized by a first reverse gate-to-drain capacitance (Crss) and the second MOS transistor is characterized by a second Crss that is greater than the first Crss and wherein the first MOS transistor has a first transistor area and the second MOS transistor has a second transistor area, the second transistor area being a fraction of the first transistor area.

2. The switch circuit of claim 1, wherein a transistor area of the second MOS transistor is between 0.1% and 15% of a total area of the common active device area.

3. The switch circuit of claim 1, wherein each of the first MOS transistor and the second MOS transistor comprises one or more trench transistors, each of the one or more trench transistors having a gate terminal formed in a trench in a semiconductor layer on the common semiconductor substrate, a body region formed in the semiconductor layer and a source region formed proximate the trench.

4. The circuit of claim 3, wherein for each trench transistor of the one or more trench transistors of the second MOS transistor includes a trench and gate terminal formed to a greater depth in the semiconductor layer than for the one or more trench transistors of the first MOS transistor.

5. The circuit of claim 3, wherein trenches for each trench transistor of the one or more trench transistors of the first and second MOS transistors are formed to a common depth and a depth of the body region formed in the semiconductor layer is greater for the first MOS transistor than for the second MOS transistor.

6. The circuit of claim 1, wherein the first MOS transistor is a first NMOS transistor and the second MOS transistor is a second NMOS transistor.

7. The switch circuit of claim 6, wherein each of the first NMOS transistor and the second NMOS transistor comprises one or more trench transistors, each of the one or more trench transistors having a gate terminal formed in a trench in an N-type semiconductor layer on the common N-type semiconductor substrate, a P-type body region formed in the semiconductor layer and an N-type source region formed proximate the trench.

8. The switch circuit of claim 6, wherein the switch circuit is a high-side switch of a buck converter.

9. The switch circuit of claim 8, wherein the buck converter includes a low-side switch having an input terminal, a output terminal, and a low-side control terminal, wherein the second terminal is connected to the input terminal.