Stress in trigate devices using complimentary gate fill materials

Abstract

Embodiments relate to an improved tri-gate device having gate metal fills, providing compressive or tensile stress upon at least a portion of the tri-gate transistor, thereby increasing the carrier mobility and operating frequency. Embodiments also contemplate method for use of the improved tri-gate device.

Description

RELATED APPLICATION

None.

FIELD

Embodiments relate to apparatus, system and method for use of CMOS devices, specifically CMOS devices with stressed channels using complimentary gate fill materials.

BACKGROUND

Metal-oxide-semiconductor field effect transistors (FETs) generally include a substrate made of a semiconductor material, such as silicon or silicon on oxide (“SOI”). The transistors typically include a source region, a channel region and a drain region within the substrate. The channel region is located between the source and the drain regions.

A tri-gate FET is a non-planar FET which may be implemented on either a bulk silicon substrate or an SOI substrate. An SOI transistor differs from a bulk transistor in that its body is not externally biased unless a specific layout is used, such as in a body-contacted transistor. A tri-gate FET uses a raised source/drain to reduce parasitic resistances. The raised channel has an added benefit of allowing for the gate to wrap around the channel. Field effects from the wrap-around gate affect the channel on three sides, thus forming a tri-gate device. The tri-gate device allows for reducing the leakage current when the transistor is in the off state. Tri-gate devices may be fabricated as either an N-type MOS (NMOS) or a P-type MOS (PMOS). Compared to planar transistors at the same off-state leakage current, the tri-gate FET can provide higher NMOS and PMOS drive current. A pair of tri-gate FETs, one NMOS and one PMOS, can be configured together to form a CMOS device. It is desirable to provide a CMOS device having a high drive current.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows a schematic of a single-fin tri-gate device;

FIG. 2 shows stress in the channel from a 1% stressed copper gate metal fill;

FIG. 3 shows the mobility response to stress for a NMOS device assuming 110 sidewall orientation for the tri-gate device;

FIG. 4 shows the mobility response to stress for a PMOS device assuming 110 sidewall orientation for the tri-gate device;

FIG. 5 shows a schematic of two different gate metal fills for N and P tri-gate devices.

DETAILED DESCRIPTION

In the following description, numerous specific details are set forth. However, embodiments may be practiced without these specific details. In other instances, well-known circuits, structures and techniques have not been shown in detail in order not to obscure the understanding of this description. The term “a” or “an” in the claims means “one or more.”

The following glossary defines terms used herein unless a different meaning is assigned within the context of usage. The Authoritative Dictionary of IEEE Standards Terms, Seventh Edition, should be referenced for terms not otherwise defined herein.

Acronym Definition
DIBL    drain induced barrier lowering
Hsi     Channel height
Lg      Gate length
MPa     Megapascal
STI     Shallow trench isolation
WSI     Width of Trigate Fin

Embodiments include a semiconductor device, having a channel raised above a substrate, the channel having a major axis; and a gate wrapped around at least a portion of the channel, the gate aligned substantially perpendicular to the major axis, the gate imparting a stress upon the channel, such that the charge carrier mobility is improved.

Optionally, in the semiconductor device, the gate further includes a gate metal fill that exerts the stress upon the channel.

Optionally, in the semiconductor device, the channel further includes an N-type material; and the gate is configured to produce tensile stress in a current flow direction.

Optionally, in the semiconductor device, the gate is made of a copper gate metal fill.

Optionally, in the semiconductor device, the channel further comprising a P-type material; and the gate is configured to produce compressive stress in a current flow direction.

Optionally, in the semiconductor device, the gate includes a tungsten gate metal fill.

Optionally, the semiconductor device further includes a shallow trench isolation layer between the N and PFET.

Embodiments include a semiconductor device, having a first channel (NFET) raised above a substrate, the first channel having a first major axis; a second channel (PFET) raised above the substrate, the second channel having a second major axis generally parallel to the first major axis, further comprising a material complementary to the material of the first channel; a first gate wrapped around at least a portion of the first channel, the first gate aligned substantially perpendicular to the first major axis, the first gate imparting a stress upon the first channel; and a second gate wrapped around at least a portion of the second channel, the second gate aligned substantially perpendicular to the second major axis, the second gate imparting a stress upon the second channel, further comprising a material complementary to the material of the first gate.

Optionally, in the semiconductor device, the first gate has a depth of 10-100 nm.

Optionally, in the semiconductor device, the second gate has a depth of 10-100 nm.

Optionally, in the semiconductor device, the stress upon the first channel is substantially unequal to the stress upon the second channel.

Embodiments include a method, including raising a channel above a substrate, channel having a major axis; wrapping a gate around at least a portion of the channel, the gate aligned substantially perpendicular to the major axis; and stressing the channel with the stress of the gate.

Optionally, in the method, the stressing improves charge carrier mobility.

Optionally, in the method, the stressing further comprising tensile stressing in current flow direction, wherein the channel comprises is an NFET.

Optionally, in the method, the stressing further includes compressive stressing in current flow direction, wherein the channel comprises a PFET.

Optionally, the method further includes isolating the two different channels (N and P), by using a shallow trench isolation layer embedded in the substrate.

As gate lengths decrease with improved generations of devices, it is increasingly difficult to fabricate the silicon MOSFET devices with reasonably low source-to-drain leakage in the off-state.

Tri-gate transistor structures address the problem of reduced leakage at off-state. The tri-gate transistor includes a thin semiconductor body formed on a substrate. The substrate can be an insulating substrate or a semiconductor substrate. A gate dielectric is formed on the top surface and the sidewalls of the semiconductor body. A gate electrode is formed on the gate dielectric on the top surface of the semiconductor body and is formed adjacent to the gate dielectric formed on the sidewalls of the semiconductor body. Source and drain regions are formed in the semiconductor body on opposite sides of the gate electrode. Because the gate electrode and the gate dielectric surround the semiconductor body on three sides, the transistor essentially has three separate channels when the channel is in the ON state. The “width” of a transistor is equal to the sum of each of the three sides of the semiconductor body. Larger “width” transistors can be formed by connecting several tri-gate transistors together via a common source and drain regions.

Because the width of the trigate transistor is smaller than the depletion width, the formation of a fully depleted transistor is possible. Hence a trigate device is characterized by a very sharp, near-ideal subthreshold slope, and a reduced drain induced barrier lowering (“DIBL”) short channel effect of less than about 100 mV/V even at short channel lengths. The lower DIBL effect results in a lower leakage current when a device using tri-gate transistor is turned “OFF”, producing lower power consumption.

The amount of current flowing through a channel which has a given electric field across it is generally proportional to the mobility of the carriers in the channel. Thus, by increasing the carrier mobility in the channel, the magnitude of current and operating speed (i.e., the operating frequency) of the transistor can be increased.

Carrier mobility within a semiconductor device substrate can be affected by, for example, mechanical stresses on and within the semiconductor device. That is, certain stresses within a semiconductor device are known to enhance semiconductor device characteristics. For example in Silicon, tensile stress improves electron mobility and degrades hole mobility in the current flow direction, and compressive stress degrades electron mobility and improves hole mobility in the current flow direction. Thus, to improve the characteristics of a semiconductor device, tensile and/or compressive stresses are created in the channel of the n-type devices (e.g., NFETs) and/or p-type devices (e.g., PFETs). The same stress component, for example tensile stress or compressive stress, improves the device characteristics of one type of device (i.e., n-type device or p-type device) while adversely affecting the characteristics of the other type device.

Related art describes adding a material layer on top of a gate electrode to induce stress in a planar MOSFET, for instance when the gate electrode is doped polysilicon. The related art describes NMOS and PMOS devices that require opposite stress states along-the-channel stress because that was relevant to the geometry described in the related art.

The embodiments described herein relate to CMOS devices based on a tri-gate architecture, using stress formed from gate-fill to affect charge carrier mobility. The embodiments focus on a non-planar tri-gate device architecture wherein at least a portion of the gate electrode material is used to generate stress. The non-planar nature of the tri-gate device and the associated wrapping-around of the gate electrode allows additional stress components to be designed for performance gain, for instance, stress normal to the gate plane and stress in the gate plane perpendicular to the current flow direction.

Enhanced carrier mobility due to mechanical stress is proportional to the amount of stress, therefore it is desirable to create as much stress in the semiconductor device substrate as possible, up to the plastic deformation limit of the material. Additionally, stresses in other structures of the device may be generated due to appropriately adjusting characteristics in the deposition process, or introducing stress-producing dopants into the deposition process. Increasing stress by adjusting the deposition process may provide a limited increase in stress. Thus, it is desirable to develop better methods of coupling stress into the channel region of a CMOS device to increase the amount of stress in the channel and enhancing carrier mobility. Applicants have calculated that compressive or tensile stress in the current flow direction may be enhanced by appropriate selection of metallic gate fill materials, i.e., the materials that make up the gate.

Intrinsically stressed gate metal fill exerts stress on the channel. Intrinsic stress is a phenomenon that develops in a thin film during the growth of the thin film. The intrinsic stress is dependent, among other things, on temperature and mechanism of the growth process. Therefore, intrinsic stress is not a property that is inherent to a thin film (e.g., a gate metal film) but rather is a property that has to be engineered into the design and fabrication of the thin film. The design and fabrication of thin films having a desired amount of intrinsic stress is well known to persons of ordinary skill in the art. As used herein, the terms “stress” or “lattice stress” refer to the same phenomenon as intrinsic stress, unless a different meaning is clearly intended, either explicitly or implicitly.

FIG. 1 shows a tri-gate geometry in which the gate 11 wraps around the channel 12, thereby exerting more stress compared to a planar structure. A gate 11 like that shown in FIG. 1 resembles a fin and may be referred to as a fin. Material stress is known to improve charge carrier mobility and hence increases the drive current produced. In a CMOS device, having both N-type and P-type regions, the stress should be applied simultaneously to both types of regions. The amount of stress needed requires metal fills of approximately 10-100 nm in depth along the top 13 and sides 14 of the tri-gate device, and furthermore the stress should be complementary. Preferably, the intrinsic stress beneficial for improving mobility is provided as tensile stress in the current flow direction for N-type devices and as compressive stress in the current flow direction for P-type devices. Gate metal fill that expands, i.e., gate metal fill made of copper, exerts the desired tensile stress upon an N-type tri-gate channel. Metal fill that shrinks, i.e. tungsten, imparts the desired compressive stress in the current flow direction to a P-type tri-gate structure. Hence in a CMOS device, embodiments use complementary metal fills. The stresses in N-type and P-type tri-gate devices may be substantially unequal, but as much stress as possible should be imparted to both N-type and P-type tri-gate devices in order to favorably increase mobility and drive current for both types of devices.

FIG. 1 shows a schematic of a single-fin tri-gate device. Persons skilled in the art will understand that embodiments are not limited to a single fin, and may be practiced using multiple-fin tri-gate devices. The source of the MOSFET is in the foreground and the drain is in the background, with the channel between the source and drain having a major axis defined as the direction between source and drain. Each MOSFET channel has one major axis; however, if a semiconductor device has a plurality of MOSFETs, then there will be a plurality of major axes in total on the semiconductor device. The source and drain have a width and height of W_SI and H_SI, respectively. The gate is shown wrapping around three sides of the junction between the source and drain. The gate length is L_G. The arrows in FIG. 1 show the directions of intrinsic stress. Stress in the direction of current flow is shown by the pair of arrows between the source and drain. Similarly, intrinsic stress applied in the “in-plane” direction for the top surface is shown by the pair of arrows pointing left and right. Intrinsic stress applied in the “in-plane” direction for the side-surface is shown by the vertical pair of arrows. The intrinsic stress applied in the “surface normal” direction for the top surface is shown by the vertical arrow. Intrinsic stress applied in the “surface normal” direction for the side surface is shown by the pair of arrows pointing up and down. Persons skilled in the art will also understand that when the tri-gate device is fabricated on an integrated circuit, shallow trench isolation (“STI”) layer may be used between N and P-type the tri-gate devices in order to prevent electrical current leakage between adjacent semiconductor device components.

The top half of FIG. 2 shows the tensile stress in the current flow direction 21 and in-plane direction 22 on the tri-gate silicon channel in a tri-gate CMOS device, averaged over the tri-gate area, as a function of the pitch between the N-type and P-type channels, and further assuming a copper metal fill that is intrinsically 1% stressed compared to the relaxed state. The compressive stress in the surface normal direction 23 on the tri-gate silicon channel in the surface normal direction is shown in the bottom half of FIG. 2, also intrinsically 1% stressed compared to the relaxed state. “W_Pitch” is the distance between individual channels, of a particular type of multi-fin device for example multi-fin N-type or multi-fin P-type devices. Increasing tensile stress is shown increasing in the upward direction in the top half of FIG. 2. Increasing compressive stress is shown increasing in the downward direction in the bottom half of FIG. 2.

FIG. 3 shows the percentage change in electron mobility in Silicon with respect to an unstressed device, as a function of stress applied to a long-channel NMOS tri-gate device (LCNMOS), assuming 110 sidewall orientation for the tri-gate device. Although the stress response is illustrated assuming (110) sidewall surface for a tri-gate LCNMOS device, persons skilled in the art will understand that the mobility enhancement is not limited to (110) surface orientations. The electric field is at 1 MV/cm. A pair of curves are presented for stress applied in each of the three directions (in-plane 31, current flow 32, and surface normal 33). Tensile stress (i.e., positive stress) is on the right side of FIG. 3, and compressive stress (i.e., negative stress) is on the left side of FIG. 3. Each pair of curves represents two calculations of the resulting stress.

FIG. 4 presents a similar increase in mobility in Silicon arising from the application of stress for PMOS. The percentage change in hole mobility with respect to an unstressed device is presented as a function of stress applied in all three dimensions (in-plane 41, current flow 42, and surface normal 43) to a long-channel PMOS tri-gate device (LCPMOS), assuming 110 sidewall orientation for the tri-gate device.

FIG. 5 shows a cross-sectional view of two tri-gate devices 10 of FIG. 1, configured side-by-side to produce for instance a CMOS device. The cross-section is taken through the gate structures 54 and 55, showing the two different gate metal fills for the N-type and P-type tri-gate devices. The source-to-drain N-type channel 51 and the source-to-drain P-type channel 52 are shown in cross-section, with current flow in a direction perpendicular to the cross-section. The STI 53 is a layer of material that prevents electrical current leakage to or from the tri-gate devices.

Other embodiments contemplate methods of fabricating the above-described embodiments, methods of using the above-described embodiments, and apparatus or systems using the above-described embodiments.

This application may disclose several numerical range limitations that support any range within the disclosed numerical ranges even though a precise range limitation is not stated verbatim in the specification because the embodiments could be practiced throughout the disclosed numerical ranges. Finally, the entire disclosure of the patents and publications referred in this application, if any, are hereby incorporated herein in entirety by reference.

Claims

1. A semiconductor device, comprising: a first fin comprising an N-type material raised above a substrate, the first fin having a first top surface, a first pair of sidewalls and a first current flow direction along a first major axis of the first fin;a second fin comprising a P-type material raised above the substrate, the second fin having a second top surface, a second pair of sidewalls and a second current flow direction along a second major axis of the second fin generally parallel to the first major axis;a first gate dielectric layer wrapped around the first top surface and the first pair of sidewalls;a first gate comprising an intrinsically compressively stressed copper metal fill, the first gate in direct contact with the first gate dielectric layer on the first top surface and on the first pair of sidewalls, the first gate aligned substantially perpendicular to the first major axis, wherein the intrinsic compressive stress of the first copper metal fill exerts a tensile stress upon the first top surface and the first pair of sidewalls in the first current flow direction along the first major axis;a second gate dielectric layer wrapped around the second top surface and the second pair of sidewalls; anda second gate comprising an intrinsically tensilely stressed tungsten metal fill, the second gate in direct contact with the second gate dielectric layer on the second top surface and on the second pair of sidewalls, the second gate aligned substantially perpendicular to the second major axis, wherein the intrinsic tensile stress of the tungsten metal fill exerts a compressive stress upon the second top surface and the second pair of sidewalls in the second current flow direction along the second major axis.

2. The semiconductor device of claim 1, wherein the first gate has a depth of 10-100 nm.

3. The semiconductor device of claim 1, wherein the second gate has a depth of 10-100 nm.

4. The semiconductor device of claim 1, wherein the tensile stress exerted upon the first channel is substantially unequal to the compressive stress exerted upon the second channel.

5. The semiconductor device of claim 1, wherein the copper metal fill is intrinsically 1% stressed compared to a relaxed state.