US 20060244045 A1
Semiconductor devices and fabrication methods are presented, in which transistor gate structures are created using doped metal silicide materials. Upper and lower metal silicides are formed above a gate dielectric, wherein the lower metal silicide is doped with n-type impurities for NMOS gates and with p-type impurities for PMOS gates, and wherein a silicon may, but need not be formed between the upper and lower metal silicides. The lower metal silicide can be deposited directly, or may be formed through reaction of deposited metal and poly-silicon, and the lower silicide can be doped by diffusion or implantation, before or after gate patterning.
35. A transistor gate structure, comprising:
a gate dielectric formed above a semiconductor body;
a first metal silicide on, and in direct physical contact with, the gate dielectric, the first metal silicide doped with substantially single polarity impurities and having a first metal component; and
a second metal silicide having a metal component different from the first metal component, above the first metal silicide.
37. The gate structure of
38. The gate structure of
39. The gate structure of
40. The gate structure of
41. The gate structure of
The present invention relates generally to semiconductor devices and more particularly to doped silicide MOS transistor gates and fabrication methods for making the same.
Field effect transistors (FETs) are widely used in the electronics industry for switching, amplification, filtering, and other tasks related to both analog and digital electrical signals. Most common among these are metal-oxide-semiconductor field-effect transistors (MOSFETs), wherein a gate contact or electrode is energized to create an electric field in a channel region of a semiconductor body, by which electrons are allowed to travel through the channel between a source region and a drain region of the semiconductor body. The source and drain regions are typically formed by adding dopants to targeted regions on either side of the channel. A gate dielectric or gate oxide is formed over the channel, and a gate electrode or gate contact is formed over the gate dielectric. The gate dielectric and gate electrode layers are then patterned to form a gate structure overlying the channel region of the substrate.
In operation of the resulting MOS transistor, the threshold voltage (Vt) is the gate voltage value required to render the channel conductive by formation of an inversion layer at the surface of the semiconductor channel. Complementary MOS (CMOS) devices have become widely used in the semiconductor industry, wherein both n-channel and p-channel (NMOS and PMOS) transistors are used to fabricate logic and other circuitry. For enhancement-mode (e.g., normally off) devices, the threshold voltage Vt is positive for NMOS and negative for PMOS transistors. The threshold voltage is dependent upon the flat-band voltage, where the flat-band voltage depends on the work function difference between the gate and the substrate materials, as well as on surface charge.
The work function of a material is a measure of the energy required to move an electron in the material outside of a material atom from the Fermi level, and is usually expressed in electron volts (eV). For CMOS products, it is desirable to provide predictable, repeatable, and stable threshold voltages for the NMOS and PMOS transistors. To establish Vt values, the work functions of the PMOS and NMOS gate contact and the corresponding channel materials are independently tuned or adjusted through gate and channel engineering, respectively.
Channel engineering typically involves shallow dopant implants to the prospective channel regions of the semiconductor body, sometimes referred to as threshold adjust (Vt adjust) implants, where the implanted impurities behave as a sheet of fixed charge located under the gate oxide. A Vt adjust implant for the NMOS devices introduces boron or other p-type impurities into the NMOS channel region to raise the channel work function (sometimes referred to as a VTN implant), and a Vt adjust implant for the PMOS devices introduces arsenic, phosphorus, or other n-type impurities to lower the PMOS channel work function (VTP implant). In this manner, the Vt for the channels can be separately adjusted for NMOS and PMOS devices. Channel engineering typically includes multiple implants, for example, a Vt adjust implant, a punch-thru implant to suppress punch-through, and a channel stop implant, for each of the NMOS and PMOS devices.
Gate engineering is employed in combination with channel engineering to adjust the work function of the gate contact materials, where different gate work function values are set for PMOS and NMOS gates. The need to independently adjust PMOS and NMOS gate work functions has made poly-silicon attractive for use as gate contact material in CMOS processes, since the work function of poly-silicon can be easily raised or lowered by doping the poly-silicon with p-type or n-type impurities, respectively. The PMOS poly-silicon gates are doped with p-type impurities and NMOS gate poly-silicon is doped with n-type dopants, typically during implantation of the respective source/drain regions following gate patterning. In this way, the final gate work functions are typically near the Si conduction band edge for NMOS and near the valence band for PMOS. The provision of dopants into the poly-silicon also has the benefit of increasing the conductivity of the gate contact. Poly-silicon has thus far been widely using in the fabrication of CMOS devices, wherein the gate engineering (e.g., implants) are conventionally tuned to provide a desired gate contact conductivity (e.g., sheet resistance value), and the threshold voltage fine tuning is achieved by tailoring the Vt adjust implants to change the channel work function.
At 32, the PMOS source/drain regions and the PMOS poly-silicon gate structures are implanted with p-type dopants to further define the PMOS source/drains, and to render the PMOS gates conductive. Similarly, the NMOS source/drain regions and the NMOS poly-silicon gate structures are implanted at 34 with n-type dopants, further defining the NMOS source/drains and to rendering the NMOS gates conductive. Thereafter, the source/drains and poly-silicon gates are silicided at 36 and back end processing (e.g., interconnect metalization, etc.) is performed at 38, before the process 10 ends at 40. In the conventional process 10, the channel engineering implants at 16 and 18 shift the work functions of the PMOS and NMOS channel regions, respectively, to compensate for the changes in the PMOS and NMOS poly-silicon gate work functions resulting from the source/drain implants at 32 and 34, respectively. In this manner, the desired work function difference between the gates and channels, and hence the desired threshold voltages, may be achieved for the resulting PMOS and NMOS transistors.
The gate dielectric or gate oxide between the channel and the gate contact is an insulator material, typically SiO2 or other dielectric, that operates to prevent large currents from flowing from the gate contact into the channel when a voltage is applied to the gate electrode. The gate dielectric also allows an applied gate voltage to establish an electric field in the channel region in a controllable manner. Continuing trends in semiconductor product manufacturing include reduction in electrical device feature sizes (scaling), as well as improvements in device performance in terms of device switching speed and power consumption. MOS transistor performance may be improved by reducing the distance between the source and the drain regions under the gate electrode of the device, known as the gate or channel length, and by reducing the thickness of the layer of gate oxide that is formed over the semiconductor surface.
However, there are electrical and physical limitations on the extent to which SiO2 gate dielectrics can be made thinner. These include gate leakage currents tunneling through the thin gate oxide, limitations on the ability to form very thin oxide films with uniform thickness, and the inability of very thin SiO2 gate dielectric layers to prevent dopant diffusion from the gate poly-silicon into the underlying channel. Accordingly, recent scaling efforts have focused on high-k dielectric materials having dielectric constants greater than that of SiO2, which can be formed in a thicker layer than scaled SiO2, and yet which produce equivalent field effect performance. A thicker high-k dielectric layer can thus be formed to avoid or mitigate tunneling leakage currents, while still achieving the required electrical performance equivalent (e.g., capacitance value) of a thinner SiO2.
Another shortcoming of scaled CMOS devices having poly-silicon gate contacts is known as poly-silicon depletion. Poly-silicon depletion occurs when annealing or other thermal back-end processing following the implants at 32 and 34 is insufficient to drive the implanted impurities down the entire depth of the poly-silicon gate structures. In this situation, a bottom portion of the poly-silicon gate contact near the gate dielectric is “depleted” of charges, and acts as an insulator. The depleted portion of the gate contact and the gate dielectric operate as series connected capacitors, resulting in a reduced effective gate capacitance, which reduces the drive current capability of the device. Consequently, poly-silicon depletion causes reduction in device performance which leads to poor unscalable devices. Simply increasing the implant energy and/or anneal time to combat poly-silicon depletion has adverse results, in that the corresponding depths of the concurrently implanted source/drain regions are increased.
With the relatively thick gate dielectrics and gate contact structures of the past, poly depletion was not critical to achieve desired device performance. However, as gate dielectrics and gate contacts continue to become smaller through scaling, the poly-silicon depletion problem is more pronounced, wherein poly-silicon depletion regions of 2 to 4 angstroms become a significant fraction of the overall effective gate capacitance. Thus, while poly-silicon gate contacts have previously offered flexibility in providing dual work functions for CMOS processes, the future viability of conventional poly-silicon gate technology is lessened as scaling efforts continue. Accordingly, attention has recently been directed again to the possibility of using metal gate contacts in CMOS products, where the metal gate materials conceivably do not need doping for conductivity improvement. Although this approach presumably avoids poly-silicon depletion issues with respect to gate capacitance, there remains a need for dual or differentiated work function capability (e.g., for PMOS and NMOS transistors) in CMOS processes. In this regard, metal work functions are not shifted as easily by the same amounts as was the case for poly-silicon. Accordingly, there is a need for improved CMOS transistor gate designs and fabrication techniques by which the benefits of scaling can be achieved while avoiding or mitigating the poly-silicon depletion degradation found in conventional devices.
The following presents a simplified summary in order to provide a basic understanding of one or more aspects of the invention. This summary is not an extensive overview of the invention, and is neither intended to identify key or critical elements of the invention, nor to delineate the scope thereof. Rather, the primary purpose of the summary is to present some concepts of the invention in a simplified form as a prelude to the more detailed description that is presented later.
The invention relates to semiconductor devices and fabrication methods, wherein transistor gate structures are created using doped metal silicide materials. A first metal silicide is formed above a gate dielectric and the silicide is doped with n-type impurities for NMOS gates or with p-type impurities for PMOS gates. The inventors have appreciated that the gate work function of the resulting NMOS and PMOS transistors can be successfully adjusted through appropriate doping the gate silicide, thereby facilitating elimination or simplification of channel engineering. The doped first metal silicide may comprise a refractory metal such as molybdenum, tungsten, tantalum, titanium, or others having a high melting temperature, and may be formed by direct silicide deposition or through reacting metal and silicon deposited above the gate dielectric. In this regard, a single starting material may be employed to form both PMOS and NMOS gate structures, thus avoiding process integration problems associated with using different gate materials. Dopants may be added to the first silicide through any means, such as direct implantation into a reacted or deposited silicide or implantation into unreacted metal and/or unreacted silicon followed by annealing to concurrently diffuse the dopants into the metal and to create the resulting metal silicide, where the dopants can be introduced before or after gate patterning.
The gate structures include a second upper metal silicide located above the doped first metal silicide. A silicon material may be provided between the first and second metal silicides, or the second metal silicide may overlie the doped first metal silicide thereby forming a silicide bilayer above the gate dielectric. In order to facilitate controlled dopant distribution within the doped first metal silicide for setting the gate work function (e.g., particularly at the gate dielectric interface), the first metal silicide may be relatively thin, wherein the provision of the upper silicide and/or the intervening silicon can advantageously reduce the gate contact sheet resistance. In one example, the first metal silicide is doped MoSi2 where poly-silicon is formed over the doped metal silicide. The poly-silicon may also be doped, for example, via source/drain or other implantations, to render the poly-silicon conductive. A second silicide is then formed over the poly-silicon, wherein the poly-silicon may be reacted with a deposited second metal (e.g., nickel) to form the upper silicide. The upper silicide and the remaining poly-silicon may thus be of lower resistivity than the first metal silicide. In another example, the entire poly-silicon is reacted (e.g., consumed) in forming the upper silicide, whereby the gate stack includes a metal silicide bilayer.
In accordance with an aspect of the invention, a method is provided for fabricating PMOS and NMOS gate structures in a semiconductor device. The method comprises forming a gate dielectric, forming an n-doped first metal silicide above the gate dielectric in an NMOS region, and forming a p-doped first metal silicide above the gate dielectric in the PMOS region. The method further comprises forming a second metal silicide above the doped first metal silicide in the NMOS and PMOS regions. The first and second metal silicides may be different, for example, where the lower silicide includes a refractory metal (e.g., molybdenum, tungsten, tantalum, titanium, etc.) and the upper silicide includes nickel, and the method may further comprise forming silicon between the first and second metal silicides in the NMOS and PMOS regions.
Another aspect of the invention provides a semiconductor device comprising NMOS and PMOS gate structures. The NMOS gate structure comprises a gate dielectric above a semiconductor body, an n-doped first metal silicide structure above the gate dielectric, and a second metal silicide above the n-doped first metal silicide. The PMOS gate structure comprises a gate dielectric above a semiconductor body, a p-doped first metal silicide structure above the gate dielectric, and a second metal silicide above the p-doped first metal silicide. In one implementation, the device further comprises silicon between the first and second metal silicides in the NMOS and PMOS gate structures. Yet another aspect of the invention provides a transistor gate structure, comprising a gate dielectric formed above a semiconductor body, a first metal silicide above the gate dielectric, the first metal silicide being doped with n or p-type impurities, and a second metal silicide above the first metal silicide.
The following description and annexed drawings set forth in detail certain illustrative aspects and implementations of the invention. These are indicative of but a few of the various ways in which the principles of the invention may be employed.
One or more implementations of the present invention will now be described with reference to the attached drawings, wherein like reference numerals are used to refer to like elements throughout, and wherein the illustrated structures are not necessarily drawn to scale. The invention relates to semiconductor devices, as well as MOS transistor gate structures and fabrication methods therefor, in which first and second metal silicides are formed in a gate structure and the first metal silicide is doped.
In accordance with an aspect of the invention, the devices 50 and 52 include MOS transistor gate structures, comprising a gate dielectric layer 62 formed above a channel region of the substrate 54, an n-doped first metal silicide 64 above the gate dielectric 62 having a thickness of about 100 Å or less, and a second metal silicide 66 above the first metal silicide 64, wherein the exemplary gate structures of
The transistor 58 a of
In the illustrated devices 50 and 52, the first and second metal silicides 64 and 66, respectively, are different, wherein the first metal silicide 64 comprises a refractory metal such as molybdenum, tungsten, tantalum, titanium, or others, while the second metal silicide comprises nickel. Alternatively, the first and second metal silicides 64 and 66 can be the same type and/or can comprise any suitable metal within the scope of the invention. However, the inventors have appreciated that the use of refractory metal silicides, such as MoSi2 or others comprising W, Ti, Ta, etc., for the first silicide 64 provides high melting temperatures to withstand fabrication processing. In this regard, nickel or other second silicides 66 provide low resistivity, wherein the second silicide 66 may be formed later in a fabrication flow, and thus only be exposed to relatively lower temperature back-end processing.
While NMOS transistors 58 are illustrated in
Another aspect of the invention provides methods for fabricating PMOS and NMOS transistors in a semiconductor device. In the examples below, an n-doped first metal silicide is formed above a gate dielectric in an NMOS region of a semiconductor device, a p-doped first metal silicide is formed above the gate dielectric in a PMOS region, and a second metal silicide is formed above the doped first metal silicide in the NMOS and PMOS regions. In accordance with the invention, the n-doped and p-doped first silicides can be formed by separate processes for the NMOS and PMOS regions, respectively. However, the invention also advantageously provides for forming the doped first silicides using a common starting material in both NMOS and PMOS regions. In one example, a single starting first metal silicide is deposited above the gate dielectric and then selectively doped in the NMOS and PMOS regions. In another implementation, a common first metal and a first silicon are deposited in both regions, and are subsequently annealed to form the first silicide, either before, or after selective doping of the NMOS and PMOS regions. Thereafter, a second or upper silicide, doped or undoped, is formed above the first silicide, and an optional silicon layer may be formed between the first and second silicides.
Although the exemplary method 100 of
The method 100 begins at 102 in
At 106, a gate dielectric is formed in the NMOS and PMOS regions using any suitable materials, material thicknesses, and processing steps, including a single thermal oxidation or deposition or combinations thereof to form a gate dielectric above the semiconductor body. The invention may be employed in conjunction with any gate dielectric material, such as SiO2, SiON, high-k dielectrics, and stacks or combinations thereof, including but not limited to binary metal oxides including aluminum oxide (Al2O3), zirconium oxide (ZrO2), hafnium oxide (HfO2), lanthanum oxide (La2O3), yttrium oxide (Y2O3), titanium oxide (TiO2), as well as their silicates and aluminates; metal oxynitrides including aluminum oxynitride (AlON), zirconium oxynitride (ZrON), hafnium oxynitride HfON), lanthanum oxynitride (LaON), yttrium oxynitride (YON), as well as their silicates and aluminates such as ZrSiON, HfSiON, LaSiON, YSiON, etc.; and perovskite-type oxides including a titanate system material such as barium titanate, strontium titanate, barium strontium titanate (BST), lead titanate, lead zirconate titanate, lead lanthanum zirconate titanate, barium lanthanum titanate, barium zirconium titanate; a niobate or tantalate system material such as lead magnesium niobate, lithium niobate, lithium tantalate, potassium niobate, strontium aluminum tantalate and potassium tantalum niobate; a tungsten-bronze system material such as barium strontium niobate, lead barium niobate, barium titanium niobate; and bilayered perovskite system material such as strontium bismuth tantalate, bismuth titanate and others. Separate processing may optionally be employed to form different gate dielectrics in the NMOS and PMOS regions within the scope of the invention. In the examples illustrated and described herein, a single thermal oxidation is performed at 106 to create a thin SiO2 or gate oxide overlying the substrate in the NMOS and PMOS regions.
Gate structures are then formed at 110, with the general gate fabrication principles being illustrated in
A silicon layer is formed at 116 above the first metal silicide in the NMOS and PMOS regions, some or all of which may subsequently be consumed in forming a second metal silicide above the first metal silicide, wherein the silicon may, but need not, be n-doped or p-doped in the NMOS and PMOS regions, respectively. Any suitable formation techniques may be employed to create the silicon at 116, for example, chemical vapor deposition (CVD) processes, etc. The silicon may be formed at 116 directly over the first metal silicide or may be formed over one or more intervening material layers within the scope of the invention. The silicon and first metal silicide are patterned at 118 to form NMOS and PMOS gate structures using any suitable techniques, for example, such as masked etching.
At 120, the gate structures are optionally encapsulated using any suitable materials and processing steps. Highly-doped drain (HDD) implants are performed at 122 to introduce p-type dopants (e.g., boron, etc.) into prospective PMOS source/drains and to provide n-type dopants (e.g., arsenic, phosphorus, etc.) to prospective NMOS source/drains, using any suitable processing steps and operational parameters, wherein the patterned gate structures and isolation structures operate as implantation masks. The HDD implants at 122 also provide n-type dopants to the silicon in the NMOS regions and p-type dopants to the silicon in the PMOS regions in the illustrated example, although the gate silicon layers can alternatively be masked during the implantation operations at 122. Sidewall spacers are formed at 124 along the lateral sidewalls of the patterned gate structures, using any suitable process techniques and materials, for example, by depositing and etching SiO2, Si3N4, etc.
Source/drain implants are then performed at 126 and 128 to further dope the PMOS and NMOS source/drain regions of the semiconductor body, wherein n-type dopants may also be implanted into the silicon in the NMOS regions and p-type dopants may be implanted into the silicon in the PMOS regions. In the illustrated implementation, a PMOS source/drain implant is performed at 126, to introduce boron or other p-type impurities into the PMOS region of the semiconductor body and into the PMOS gate silicon with the NMOS region masked. Any suitable implantation technique, operational settings, and implant species may be employed at 126 to render the PMOS gate silicon conductive and to further define the PMOS source/drains in the semiconductor body.
An NMOS source/drain implant is performed at 128 to introduce arsenic, phosphorus, or other suitable n-type impurities to the NMOS region of the semiconductor body and to the NMOS gate silicon with the PMOS region masked. Any suitable implantation technique, operational settings, and implant species may be employed at 128 to render the NMOS gate silicon conductive and to further define the NMOS source/drains, wherein the implants at 126 and 128 may be performed in any order. In addition, activation anneal operations may be performed following one or both of the source/drain implants 126, 128, wherein such annealing may also operate to diffuse dopants in the first silicide.
A self-aligned silicide process is performed at 130 to silicide the source/drains and the gate silicon, wherein silicidation of the gate silicon creates a second metal silicide above the first metal silicide of the gate stacks in the exemplary implementation. In one example, nickel or other metal is deposited at 130 over the gate silicon and over the semiconductor body in the prospective source/drain regions, and the wafer is annealed to react the gate silicon and the source/drain region silicon of the semiconductor body with the deposited nickel to form nickel silicide. By this technique, the second silicide of the gate stack may be formed above the doped first metal silicide concurrently with formation of source/drain silicide contacts, although separate silicide formation techniques (e.g., reaction or deposition and etching) may be employed to form the second metal silicide of the gate structures. In one implementation (e.g.,
The second silicide can be formed anywhere above the first metal silicide of the gate structure within the scope of the invention. For example, the second silicide may be formed directly over the first metal silicide, directly over an intervening silicon, or one or more other layers may be situated between the second silicide and the first silicide or the silicon. As discussed above, the second silicide may be the same or may comprise different metal than the first silicide. For example, the second silicide may advantageously comprise nickel (e.g., low resistivity, lower melting temperature), where the first metal silicide comprises a refractory metal (e.g., molybdenum, tungsten, tantalum, titanium, etc.) having a higher melting temperature. In this manner, the second silicide may provide improved (e.g., lower) gate sheet resistance, whereas the first silicide is able to withstand relatively higher processing thermal budget between the first silicide formation at 112 and the back-end processing at 132. In addition, the provision of the intervening silicon over the first metal silicide (e.g., whether fully consumed at 130 or not) may advantageously protect the first metal silicide during fabrication processing. Back end processing (e.g., interconnect metalization, etc.) is performed at 132, before the exemplary method 100 ends at 134.
Referring now to
As illustrated in
In the exemplary device 302, the first metal silicide 312 is formed through deposition (e.g., “DEP” at 200). In
A layer of silicon 324 is then deposited in
The mask 321 is then removed and a mask 320 is formed in
In the above example of
This reacted silicide can then be provided with dopants through diffusion from an overlying doped poly-silicon (e.g., at 206-212 as described above), or be directly implanted prior to poly-silicon formation. In this case (“DOPE FIRST” at 204), the first metal silicide of the NMOS region is implanted at 216 with n-type dopants and the first metal silicide of the PMOS region is implanted with p-type dopants at 218, using appropriate masks. Thereafter an anneal may optionally be performed to diffuse the implanted dopants, and poly-silicon is deposited over the doped first metal silicide at 220. As one alternative, the silicon deposition at 220 (e.g., CVD deposition) may provide suitable thermal heating to diffuse the dopants to the desired depth, concentration, profile, etc. within the first metal silicide, whereby an extra (e.g., dedicated) anneal may not be required. Thereafter, the gate structures are patterned at 214 as described above. In the various implementations illustrated and described herein, the gate structures can alternatively be patterned at other points in the fabrication process, for example, prior to implantation and/or diffusion of dopants into the first metal silicide. Other possible alternatives include forming poly-silicon over undoped first metal silicide (reacted or deposited), and implanting dopants through the poly-silicon and into the underlying first metal silicide, with or without subsequent diffusion annealing.
Referring now to
The device 502 (
Doping of the ultimately formed first metal silicide may be accomplished using any suitable technique within the scope of the invention, including direct implantation or diffusion (“DIRECT IMPLANT” or “DIFFUSION” at 406 in
A mask 528 is formed in
Referring again to
Although the invention has been illustrated and described with respect to one or more implementations, alterations, and/or modifications may be made to the illustrated examples without departing from the spirit and scope of the appended claims. In particular regard to the various functions performed by the above described components or structures (assemblies, devices, circuits, systems, etc.), the terms (including a reference to a “means”) used to describe such components are intended to correspond, unless otherwise indicated, to any component or structure which performs the specified function of the described component (e.g., that is functionally equivalent), even though not structurally equivalent to the disclosed structure which performs the function in the herein illustrated exemplary implementations of the invention. In addition, while a particular feature of the invention may have been disclosed with respect to only one of several implementations, such feature may be combined with one or more other features of the other implementations as may be desired and advantageous for any given or particular application. Furthermore, to the extent that the terms “including”, “includes”, “having”, “has”, “with”, or variants thereof are used in either the detailed description and the claims, such terms are intended to be inclusive in a manner similar to the term “comprising”.