阅读说明：本技术 集成电路 (Integrated circuit with a plurality of transistors ) 是由 叶凌彦 萧孟轩 孙元成 于 2019-06-26 设计创作，主要内容包括：本揭示是关于一种集成电路。所揭示的技术在沟槽内局部地形成磊晶层,此沟槽具有在沟槽的侧壁中堆叠的角形凹陷。控制凹陷大小以控制在沟槽内形成的磊晶层的厚度。凹陷由覆盖层覆盖并且从最底部凹陷开始相继地逐个暴露出。磊晶层在沟槽内逐个形成,磊晶层的晶面边缘部分对准到相应凹陷中,此凹陷是为了磊晶层而相继地暴露的凹陷。(The present disclosure relates to an integrated circuit. The disclosed technique forms an epitaxial layer locally within a trench having a stacked angular recess in a sidewall of the trench. The recess size is controlled to control the thickness of the epitaxial layer formed in the trench. The recesses are covered by a cover layer and successively exposed one after the other starting from the bottommost recess. The epitaxial layers are formed one by one within the trenches with edge portions of the crystal planes of the epitaxial layers aligned in respective recesses that are successively exposed for the epitaxial layers.)

1. An integrated circuit, comprising:

a substrate;

a first device including a first stack of separate nanowire structures of a first semiconductor material over the substrate and a first gate structure surrounding the first stack of separate nanowire structures;

a second device comprising a second stack of discrete nanowire structures of a second semiconductor material over the substrate and a second gate structure surrounding the second stack of discrete nanowire structures; and

a first insulating structure located one or more of laterally between the first device and the second device, or longitudinally between the substrate and the first device.

Technical Field

The present disclosure relates to an integrated circuit.

Background

Complementary Metal Oxide Semiconductor (CMOS) transistors are building blocks for integrated circuits. Faster CMOS switching speeds require higher drive currents, which drives the gate length of CMOS transistors to continue to shrink. The shorter gate length results in an undesirable "short channel effect" in which the current control function of the gate is compromised. FinFET transistors have been developed to overcome short channel effects. As a step to further improve electrostatic control of the channel, transistors have been developed with wrap-around gates, where the gate portions may surround the semiconductor channel or channel strip from the top surface, bottom surface, and sidewalls.

Disclosure of Invention

In one embodiment, an integrated circuit includes a substrate, a first device, a second device, and a first insulating structure. The first device includes a first stack of separate nanowire structures of a first semiconductor material over the substrate and a first gate structure surrounding the first stack of separate nanowire structures. The second device includes a second stack of isolated nanowire structures of a second semiconductor material over the substrate and a second gate structure surrounding the second stack of isolated nanowire structures. The first insulating structure is one or more of laterally between the first device and the second device, or longitudinally between the substrate and the first device.

Drawings

Aspects of the present disclosure are best understood from the following detailed description when read with the accompanying drawing figures. In the drawings, like reference numerals identify similar elements or acts unless context dictates otherwise. The sizes and relative positions of elements in the drawings are not necessarily drawn to scale. In fact, the dimensions of the various features may be arbitrarily increased or reduced for clarity of discussion.

FIG. 1 is a cross-sectional view of an example Integrated Circuit (IC);

FIG. 1-1 is an enlarged view of a portion of FIG. 1;

FIG. 2 is a flow chart of an exemplary process; and

fig. 3A-3M are cross-sectional views of a wafer at various stages of an example process of fabricating a nanowire stack.

[ notation ] to show

100 Integrated Circuit Structure (IC)

102 substrate

104N type doped region (N well)

106P type doped region (P well)

110 first device

112 insulating main body

113 insulating layer

114 inclined side wall

115 space

116 is recessed

116a are recessed

116b are recessed

118 upper surface

120 separating nanowire strips

130 grid

132 outer spacer

134 internal spacer

140 source/drain regions

150 second device

152 insulating body

153 insulating layer

154 sloped side wall

156 depression

156a is recessed

156b are recessed

158 upper surface

160 separating nanowire strips

170 grid

172 outer spacer

174 internal spacer

180 source/drain regions

200 process

205 operation

210 operation

215 operation

220 operation

225 operation

230 operation

235 operation

240 operation

250 operation

255 operation

300 wafer

308 upper surface of the container

308C upper surface part

310 sacrificial substrate region

312 groove

312(1) first trench dielectric portion

312(2) alternating sections

312(3) Trench dielectric portion

312(4) alternating sections

312C upper part

314 embankment structure

314C corner portion

316 side wall

316(1) first sidewall portion

316C side wall section

320 cover layer

330(1) recess

330(2) recess

330(3) recess

330(4) recess

332 plane of crystallization

332(4) upper edge

334 crystal plane

336 ridge interface

340(1) first cover layer

340(2) coating layer

340(3) coating layer

340(4) covering layer

350 epitaxial layer

350(1) epitaxial layer

350(2) epitaxial layer

350(3) epitaxial layer

350(4) epitaxial layer

350R crystal face edge part

352(1) edge of crystal plane

352(2) edge portion of crystal plane

352(3) edge portion of crystal plane

352,352(1) edge portion of crystal plane

352,352(2) edge portion of crystal plane

352,352(3) edge portion of crystal plane

352,352(4) edge portion of crystal plane

354(1) upper surface

358 hard mask layer

360 dielectric layer

362 upper surface

370 nanowire stack

370(1) silicon nanowire strip

370(2) silicon germanium nanowire strips

370(3) silicon nanowire strips

370(4) silicon germanium nanowire strips

410 sacrificial substrate region

412 trench

414 embankment structure

416 inclined side wall

460 insulating layer

470 nanowire stack

470(1) silicon germanium nanowire strips

470(2) silicon nanowire strip

470(3) silicon germanium nanowire strips

470(4) silicon nanowire strip

Angle theta 1

Angle theta 2

Angle theta 3

Angle theta 4

Detailed Description

This disclosure describes techniques for forming gate-all-around devices from stacks of partially formed nanowire semiconductor strips. The inventors have observed that edge portions of the epitaxial layer may degrade the overall quality of the epitaxial layer during local growth of the epitaxial layer in defined spaces, such as trenches. For example, the edge portions of the crystal plane of the lower epitaxial layer cause the upper epitaxial layer to form behind the edge portions of the crystal plane of the lower epitaxial layer that combine with the edge portions of the crystal plane of the upper epitaxial layer to cause further epitaxial growth defects. The disclosed technique forms an epitaxial layer within a trench having a stacked angular recess in a sidewall of the trench. The recess size is selected to control the thickness of the epitaxial layer formed in the trench. The recesses are covered by a cover layer and successively exposed one after the other starting from the bottommost recess. The epitaxial layers are formed one after the other within the trenches with their edge portions aligned into respective recesses, which are recesses to which the epitaxial layers are successively exposed. The epitaxial process includes forming etch features to limit the epitaxial layer from vertically exceeding the exposed recess. Therefore, an epitaxial layer is not formed in the vicinity of the capping layer. Epitaxial formation also includes dynamic feedback control of the thickness to enhance the accuracy of the epitaxial layer thickness. The resulting epitaxial layers each include a facet edge portion aligned into a respective recess and extending substantially flat throughout the trench and having a desired thickness. The locally formed epitaxial layer is patterned to form a nanowire stack. A gate full-loop (GAA) transistor is formed from a partially formed nanowire stack. With the epitaxial layer formed locally, the thickness of the stacked nanowire strips is optimized locally and independently for each type of transistor.

The following disclosure provides many different embodiments, or examples, for implementing different features of the described subject matter. Specific examples of components and arrangements are described below to simplify the present description. Of course, these are merely examples and are not intended to be limiting. For example, forming a first feature over or on a second feature in the description that follows may include embodiments in which the first and second features are formed in direct contact, and may also include embodiments in which additional features may be formed between the first and second features such that the first and second features may not be in direct contact. Additionally, the present disclosure may repeat reference numerals and/or letters in the various examples. This repetition is for the purpose of simplicity and clarity and does not in itself dictate a relationship between the various embodiments and/or configurations discussed.

Furthermore, spatially relative terms, such as "below … …," "below … …," "lower," "above … …," "upper," and the like, may be used herein to describe one element or feature's relationship to another element (or elements) or feature (or features) as illustrated in the figures for ease of description. Spatially relative terms are intended to encompass different orientations of the device in use or operation in addition to the orientation depicted in the figures. The apparatus may be otherwise oriented (rotated 90 degrees or at other orientations) and the spatially relative descriptors used herein interpreted accordingly.

In the following description, certain specific details are set forth in order to provide a thorough understanding of various embodiments of the present disclosure. However, it will be understood by those skilled in the art that the present disclosure may be practiced without these specific details. In other instances, well-known structures associated with electronic components and manufacturing techniques have not been described in detail in order to avoid unnecessarily obscuring the description of the embodiments of the present disclosure.

Throughout the specification and the claims which follow, unless the context requires otherwise, the word "comprise", and variations such as "comprises" and "comprising", will be interpreted in an open-ended sense, that is, in an "including, but not limited to".

Use of ordinals (such as first, second and third) does not necessarily imply a ranking meaning for an order, but may merely distinguish between multiple instances of an action or structure.

Reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Thus, the appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

As used in this specification and the appended claims, the singular forms "a", "an" and "the" include plural referents unless the content clearly dictates otherwise. It should also be noted that the term "or" is generally employed in its sense including "and/or" unless the context clearly dictates otherwise.

The gate all-around (GAA) transistor structure may be patterned by any suitable method. For example, the structure may be patterned using one or more photolithography processes, including double patterning or multiple patterning processes. Typically, double patterning or multiple patterning processes combine photolithography and self-alignment processes, allowing for the generation of patterns with pitches that are smaller, for example, than pitches that could otherwise be obtained using a single, direct photolithography process. For example, in one embodiment, a sacrificial layer is formed over a substrate and patterned using a photolithography process. The spacers are formed using a self-aligned process with the patterned sacrificial layer. The sacrificial layer is then removed, and the remaining spacers can then be used to pattern the GAA structure.

Fig. 1 is a cross-sectional view of an example integrated circuit structure (IC)100 including CMOS components. Referring to fig. 1, an IC 100 includes a substrate 102, the substrate 102 including an N-doped region (N-well) 104 and a P-doped region (P-well) 106. A first device 110 (e.g., PMOS) is formed within and over N-well 104. A second device 150, e.g., NMOS, is formed within and over the P-well 106. One or both of the first device 110 and the second device 150 are near the respective insulating bodies 112, 152 in the respective N-wells 104, P-wells 106. Fig. 1 illustrates the insulative bodies 112, 152 separated from one another; however, embodiments of the present disclosure are not limited to the insulative bodies 112, 152 being separate from each other. For example, in other embodiments, the insulative bodies 112, 152 may contact each other. In other embodiments, one or more of the insulating bodies 112, 152 may be semiconductor bodies of various semiconductor materials. The layers 113, 153 are adjacent to the insulating bodies 112, 152, respectively, and comprise the same dielectric material as the insulating bodies 112, 152 (i.e., the layers 113, 153 may be integral parts of the respective insulating bodies 112, 152), or the layers 113, 153 may be different layers from the insulating bodies 112, 152.

Furthermore, the insulating bodies 112, 152 may be the same layer of dielectric material, e.g., the insulating bodies 112, 152 may be formed by the same deposition process of the same dielectric material. The insulative bodies 112, 152 are labeled with different numbers to indicate that such insulative bodies are associated with the first device 110 or the second device 150, respectively. In other embodiments, the insulating bodies 112, 152 are formed by different processes. The N-well 104 and P-well 106 are optional. In some embodiments, one or more of the devices 110, 150 are not formed within the doped regions 104, 106, but are formed directly in the semiconductor substrate 102, due to the insulating layers 112, 152, 113, 153 blocking movement of charge carriers.

The first device 110 and the second device 150 each include a vertical stack of multiple (two for illustration) separate nanowire strips 120, 160 and gates 130, 170 surrounding the respective separate nanowire strips 120, 160. The separate nanowire strips 120, 160 are configured as channel regions of the devices 110, 150 and form junctions with the respective source/drain regions 140, 180. The source/drain regions 140, 180 are separated from the respective gates 130, 170 by the inner spacers 134, 174 and/or the outer spacers 132, 172.

In one embodiment, the nanowire strips 120 are silicon germanium or other suitable semiconductor material. The source/drain regions 140 are doped silicon germanium or other suitable semiconductor material. The dopant is an element in group iii such as boron, gallium, indium, and the like. The nanowire strips 160 are silicon or other suitable semiconductor material. Source/drain regions 180 are doped silicon or other suitable semiconductor material. The dopant is an element in group V, such as arsenic, phosphorus, and the like.

Fig. 1 illustrates two embodiments of source/drain structures 140, 180. In the embodiment illustrated with respect to source/drain structures 140, source/drain structures 140 contact nanowire strip 120 through an edge surface of nanowire strip 120. Furthermore, the source/drain structures 140 are patterned to have spaces 115 between the source/drain structures 140 and the sloped sidewalls 114 of the insulating body 112. In the embodiment illustrated with respect to the source/drain structure 180, the source/drain structure 180 extends all the way to the insulating body 152, or specifically into the recess 156 on the sloped sidewall 154. Other structural configurations of the source/drain structures 140/180 and the semiconductor strips 120/160 are also possible and are included in the present disclosure. For example, the source/drain structures 140/180 may surround at least some of the nanowire strips 120/160. The nanowire strips 120, 160 may be retracted (as shown at 120) or extend all the way between the recesses 156 of the insulating body 152 (and/or the recesses 116 of the insulating body 112).

Fig. 1 illustrates an illustrative example of source/drain structures 140, 180 that do not limit the scope of the present disclosure. Other embodiments of the source/drain structures 140, 180 are also possible and are included in the present disclosure.

The substrate 102 may include a silicon substrate and/or other elemental semiconductor, such as germanium, in a crystalline structure. Alternatively or additionally, the substrate 102 may include a compound semiconductor such as silicon carbide, gallium arsenide, indium arsenide, and/or indium phosphide. In addition, the substrate 102 may also include a silicon-on-insulator (SOI) structure. The substrate 102 may include epitaxial layers and/or may be strained for performance enhancement purposes.

Each of the gate structures 130, 170 is formed as a replacement metal gate. The following description lists examples of materials for the gate structures 130, 170. The gate electrodes of the gates 130, 170 each comprise a conductive material, such as a metal or metal compound. Suitable metal materials for the gate electrodes of gate structures 130, 170 include ruthenium, palladium, platinum, tungsten, cobalt, nickel, and/or conductive metal oxides, as well as other suitable P-type metal materials, and include hafnium (Hf), zirconium (Zr), titanium (Ti), tantalum (Ta), aluminum (Al), aluminides, and/or conductive metal carbides (e.g., hafnium carbide, zirconium carbide, titanium carbide, and aluminum carbide), as well as other suitable materials for N-type metal materials. In some examples, the gate electrodes of the gate structures 130, 170 each include a work function layer tailored to have an appropriate work function for enhanced performance of the field effect transistor device. For example, suitable N-type work function metals include Ta, TiAl, TiAlN, TaCN, other N-type work function metals, or combinations thereof, and suitable P-type work function metal materials include TiN, TaN, other P-type work function metals, or combinations thereof. In some examples, a conductive layer (such as an aluminum layer, a copper layer, a cobalt layer, or a tungsten layer) is formed over the work function layer such that the gate electrodes of the gate structures 130, 170 each include a work function layer disposed over the gate dielectric and a conductive layer disposed over the work function layer and under the gate cap (not shown for simplicity). In one example, the gate electrodes of the gate structures 130, 170 each have a thickness in a range from about 5nm to about 40nm, depending on design requirements.

In an exemplary embodiment, the gate dielectric layer comprises an interfacial silicon oxide layer (not separately shown for simplicity), e.g., having a range of about 5 to about 10 angstroms

Thermal or chemical oxide of thickness (v). In example embodiments, the gate dielectric layer further comprises a high dielectric constant (high-k) dielectric material selected from one or more of: hafnium oxide (HfO)₂) Hafnium silicon oxide (HfSiO), hafnium silicon oxynitride (HfSiON), hafnium tantalum oxide (HfTaO), hafnium titanium oxide (HfTiO), hafnium zirconium oxide (HfZrO), combinations thereof, and/or other suitable materials. In some applications, the high-k dielectric material includes a dielectric constant (k) value greater than 6. Dielectric materials having a dielectric constant (k) value of 7 or higher are used depending on design requirements. The high dielectric constant dielectric layer may be formed by Atomic Layer Deposition (ALD) or other suitable techniques. According to embodiments described herein, the high-k dielectric layer of the gate dielectric layer comprises a range of about 10 to about 30 angstroms

Or other suitable thickness.

In one embodiment, the outer spacers 132, 172 are formed of a low-k dielectric material, such as silicon oxynitride (SiON), silicon monoxide (SiO), silicon oxycarbide (SiONC), silicon oxycarbide (SiOC), vacuum, and other dielectrics or other suitable materials. The outer spacers 132, 172 may be formed by Chemical Vapor Deposition (CVD), high density plasma CVD, spin coating, sputtering, or other suitable methods.

In one embodiment, the inner spacers 134, 174 are formed of a low-k dielectric material (e.g., a dielectric constant value less than 6). The low-k dielectric material of the inner spacers 134, 174 may have a dielectric constant that is different from the dielectric constant of the corresponding outer spacers 132, 172. The low dielectric constant material of the inner spacers 134, 174 may include one or more of the following: silicon monoxide (SiO), silicon oxycarbonitride (SiONC), silicon oxycarbide (SiOC), silicon carbide (SiC), or other suitable low-k dielectric material. In one embodiment, the dielectric constant of the material of the inner spacers 134, 174 is the same as the dielectric constant of the material of the outer spacers 132, 172. In one example, the inner spacers 134, 174 also include one or more air gaps near one or more of the respective gate structures 130, 170 or the respective source/drain structures 140, 180.

The insulating or semiconductor bodies 112, 152 each include generally sloped sidewalls 114, 154 (illustrated in phantom). In an embodiment, the angles θ 1, θ 2 between the sloped sidewalls 114, 154 and the upper surfaces 118, 158 of the insulating or semiconductor bodies 112, 152 are between about 65 degrees to about 89 degrees or other angles as described herein. The angles θ 1, θ 2 may be substantially the same or may have different values. A plurality of recesses 116, 156 are formed vertically stacked along the sidewalls 114, 154 of the insulating or semiconductor bodies 112, 152. The recess 116 on the insulating or semiconductor body 112 may be a different size than the recess 156 on the insulating or semiconductor body 152. Furthermore, two immediately adjacent recesses 116a, 116b on the sloped sidewall 114 or recesses 156a, 156b on the sloped sidewall 154 may have different sizes. As shown in fig. 1-1 (an enlarged view of the recesses 116, 156), the dimensions of the recesses 116, 156 include a thickness T and a depth D. The thickness, depth, or both of the two recesses 116, 156 may be different.

The insulating body 152 may comprise the same or different dielectric material as the insulating body 112. Semiconductor body 112/152 may include the same or different semiconductor material as substrate 102.

In other embodiments, the insulating body 112 and the insulating body 152 may both be insulating bodies of the same or different dielectric materials or both be semiconductor bodies of the same or different semiconductor materials.

In another embodiment, the insulating bodies 112, 152 may be replaced with a semiconductor material that is the same as or different from the semiconductor material of the nanowire strips 120, 160.

The nanowire ribbon 120 and the nanowire ribbon 160 are formed from a stack of epitaxial layers, each formed by local epitaxial growth. Nanowire strip 120 and nanowire strip 160 can comprise different thicknesses. Furthermore, because the sacrificial strip used in fabricating one of the devices 110, 150 and the nanowire strips 120, 160 of the other of the first and second devices 110, 150 are not fabricated from the same epitaxial layer, the internal spacers 134, 174 of one of the first and second devices 110, 150 may not be laterally aligned with the nanowire strips 160, 120 of the other of the first and second devices 110, 150. Note again that the following will be described in more detail: the nanowire ribbon 120 and the nanowire ribbon 160 are formed from a stack of epitaxial layers, each formed in a local epitaxial growth.

Fig. 1 illustrates that the insulating bodies 112, 152 completely separate the first device 110 and the second device 150 from the substrate 102, which is not limiting. In other embodiments, one or more of the first device 110 and the second device 150 may extend through the insulative bodies 112, 152 toward the substrate 102.

Fig. 1 illustrates that the first device 110 and the second device 150 are each partially embedded in the respective N-well 104 or P-well 106 of the substrate 102, which is not limiting.

Moreover, the present disclosure includes that one of the first device 110 and the second device 150 may be formed using other techniques. For example, one of the first device 110 or the second device 150 may not include a respective recess 116, 156 in a sidewall of the insulative body.

Fig. 1 illustrates the use of a dual-tub process, i.e., the formation of both P-wells 106 and N-wells 104 in a substrate 102. Other processes, such as a P-well process in an N-type substrate or an N-well process in a P-type substrate, are also possible and are included in the present disclosure. That is, one of the insulating bodies 112, 152 may be embedded in a separately doped substrate region (P-well or N-well) and the other of the insulating bodies 112, 152 is embedded in the original substrate 102 (N-type or P-type).

In an embodiment, the spaces 113, 153 between the first means 110, the second means 150 and the respective insulating/semiconductor bodies 112, 152 may be voids, air gaps, may be completely or partially filled with a dielectric material, or may be completely or partially filled with a semiconductor material. The dielectric material in the spaces 113, 153 may be the same dielectric material as the respective insulating bodies 112 or may be a different dielectric material. Furthermore, the spaces 113, 153 may be at least partially filled with a semiconductor material. The semiconductor material in the spaces 113, 153 may be the same semiconductor material as the respective source/drain structures 140, 180 or may be a different semiconductor material. Furthermore, some of the recesses 116, 156 may be at least partially filled with a semiconductor material similar to that of the nanowire strips 120, 160.

In other embodiments, as illustrated in fig. 1 with respect to the first and second devices 110, 150, respectively, the spaces 113, 153 are absent and the recesses 116, 156 are each near the semiconductor nanowire strips 120, 160, the respective source/drain structures 140, 180, or other nanowire strips.

Fig. 2 illustrates an example process of manufacturing an IC device (e.g., the example IC 100 of fig. 1). Fig. 3A-3M illustrate in cross-sectional views different stages of a wafer 300 in a process of manufacturing the example IC 100 according to the example process of fig. 2.

In an example operation 205, a wafer 300 is received. Fig. 3A illustrates that wafer 300 includes substrate 102, substrate 102 including N-wells 104 and P-wells 106 coplanar with one another at an upper surface 308 of wafer 300. As an illustrative example, the substrate 102 is a silicon substrate. FIG. 3A illustrates, as a non-limiting illustrative example, where N-wells 104 and P-wells 106 are butted against each other. N-well 104 may be doped with dopants/impurities of group V elements (e.g., arsenic, phosphorous, etc.), or various combinations thereof, in various ways. P-well 106 may be doped with dopants/impurities of group iii elements (e.g., boron, gallium, etc.), or various combinations thereof, in various ways.

Referring back to fig. 2, and also to fig. 3B, in an exemplary operation 210, sacrificial substrate regions 310, 410 are formed within one or more of the N-wells 104 and P-wells 106 using Reduced Pressure Chemical Vapor Deposition (RP-CVD) or any other suitable method. The sacrificial substrate regions 310, 410 each comprise a different material than the substrate 102. For example, in the illustrative example, in the case where the substrate 102 is silicon, the sacrificial substrate regions 310, 410 are germanium.

In another embodiment, no sacrificial substrate region is formed and subsequent processing is alternatively performed with one or more of the N-wells 104 or P-wells 106.

In another embodiment, Shallow Trench Isolation (STI) regions are formed between and/or within the N-wells 104 or P-wells 106, and the sacrificial substrate regions 310, 410 (or the original N-wells 104, P-wells 106) are subsequently formed as ring-shaped regions surrounding the STI regions in the N-wells 104 or P-wells 106.

Referring back to fig. 2, also referring to fig. 3C, in an example operation 215, trenches 312, 412 embedded in each of the sacrificial substrates 310, 410 and filled with a different material than the sacrificial substrates 310, 410 are formed. For example, the trenches 312, 412 are filled with a dielectric material (e.g., silicon oxide). In the description herein, reference numerals 312, 412 also refer to the dielectric material filled in the trenches 312, 412. In the case of forming the trenches 312, 412, the remaining portions of the sacrificial substrate 310, 410 surrounding the respective trenches 312, 412 are referred to as bank structures 314, 414, respectively.

The bank structures 314, 414 each include a sloped sidewall 316, 416. The angles θ 3, θ 4 between the sloped sidewalls 316, 416 of the bank structures 314, 414 and the upper surface 308 are between about 65 degrees to about 89 degrees. In addition, the angles θ 3, θ 4 are determined by predetermined crystal plane angles of the epitaxial layers formed in the trenches 312, 412. Specifically, the grooves θ 3, θ 4 are selected in the following manner: the respective sloped sidewalls 316, 416 are adapted to form a recess that matches the epitaxial layer's crystal plane angle. The significance of this angular range is further described herein.

Fig. 3C illustrates that the trenches 312, 412 do not extend down through the sacrificial substrate regions 310, 410 and do not reach the substrate 102, or more specifically, the N-well 104 or the P-well 106. In other embodiments, the trenches 312, 412 each extend through the respective sacrificial substrate region 310, 410 and reach the substrate 102, or more specifically, the N-well 104 or the P-well 106.

As described herein, the trenches 312, 412 may be formed before the sacrificial substrate regions 310, 410, without altering the scope of the present disclosure. Furthermore, the trenches 312, 412 may be formed directly within the N-well 104, P-well 106 without forming the sacrificial substrate regions 310, 410 before or after it. In the description herein, the sacrificial regions 310, 410 are used as illustrative examples to illustrate an example process. Similar processes may be performed directly with the trenches 312, 412 surrounded by the N-well 104, P-well 106 material.

Referring back to fig. 2, and also to fig. 3D, in an example operation 220, an overburden 320 is formed in the other sacrificial substrate 310 covering the corner portions 314C of the bank structures 314 with one of the sacrificial substrate regions (e.g., here sacrificial substrate region 410) covered by a hard mask layer 358, such as silicon carbide. As shown in fig. 3D, corner portion 314C includes portion 316C of sidewall 316 and portion 308C of upper surface 308. The capping layer 320 covers the sidewall portions 316C and the upper surface portions 308C of the corner portions 314C. In one embodiment, a pad oxide layer 322 is formed between the capping layer 320 and the upper surface portion 308C of the corner portion 314C. The pad oxide 322 may protect the corner portion 314C during the process of removing the capping layer 320. Capping layer 320 is silicon nitride or other suitable dielectric material.

An upper portion 312C (illustrated in dashed lines) of the trench material 312 is removed when the capping layer 320 is formed.

In some of the figures described below, only the N-well 104 portion of the wafer 300 is illustrated for simplicity. It will be appreciated that similar descriptions apply to the P-well 106 portion of the wafer 300 and may be performed in separate processes.

Referring back to fig. 2, and also to fig. 3E, in an example operation 225, the first sidewall portions 316(1) are exposed by removing the first trench dielectric portions 312(1) of the dielectric material of the trenches 312. In one embodiment, thickness t1 of removed first trench dielectric portion 312(1) is controlled to be substantially equal to a thickness of an epitaxial layer desired to be formed in the space of removed first trench dielectric portion 312 (1).

Referring back to fig. 2, and also to fig. 3F, in an example operation 230, a first recess 330(1) is formed through the exposed first sidewall portion 316 (1). In one embodiment, the first recess 330(1) has an angular profile and is formed with an etchant having different etch rates among different crystallographic planes/planes of the sacrificial substrate 310 material (here, germanium as an illustrative example). An angular profile refers to two planes that meet at a ridge interface. As illustrated in the enlarged view of FIG. 3F-1, the angular profile includes two flat surfaces 334 that meet at a ridge interface 336. If the etchant has an anisotropic etch rate along the different crystallographic planes of the exposed sidewall portions 316(1), a wet etch process or a reactive ion etch process may be used for the crystallographic anisotropic etch. More specifically, as illustrated in the enlarged view of FIG. 3F-1, the crystalline planes 334 have a lower etch rate than the crystalline planes 332 (illustrated in phantom). Since the angle θ 3 (fig. 3C) between about 65 to about 89 degrees is quite close to 90 degrees, the sidewalls 316 are relatively aligned with the crystallographic planes 332, and appropriate angular recesses 330(1) are formed based on the etch rate differences. A further smaller angle θ 3 (e.g., less than 65 degrees) may hinder the formation of the angular recess 330 (1).

The profile of the angular recesses 330(1) is controlled to conform to or be in line with the edge portions of the angular crystal planes of the epitaxial layers formed within the trenches 312 and between the first recesses 330 (1).

It is also contemplated that the first recess 330(1) may have other contours (e.g., semi-elliptical), as are included in the present disclosure. Furthermore, depending on the etch timing, the etch may also terminate before the high etch rate crystallographic planes 332 decrease to reach the ridge interface 336. That is, as shown in FIG. 3F-2, the first recess 330(1) may include a trapezoidal profile.

Referring back to fig. 2, and also to fig. 3G, in an example operation 235, a first capping layer 340(1) is formed covering the first recess 330 (1). In one embodiment, the first capping layer 340(1) comprises a dielectric material different from the dielectric material of the trench 312, such that the first capping layer 340(1) and the dielectric material of the trench 312 comprise different etch selectivities from each other. In one embodiment, capping layer 340(1) comprises the same dielectric material as capping layer 320, e.g., SiN here. As shown in fig. 3G, the first capping layer 340(1) also covers the sidewall portions 316C of the corner portions 314C, thereby accumulating over the capping layer 320.

Referring to fig. 3H, the sequence of operations 225, 230, 235 is repeated a predetermined number of times such that a vertical stack of a plurality of recesses 330 (illustrated as four recesses 330(1), 330(2), 330(3), and 330(4)) is formed along the sidewalls 316 of the bank structures 314. In one embodiment, two immediately adjacent removed trench dielectric portions 312 (e.g., 312(2) and 312(3)) and corresponding exposed sidewall portions comprise different thicknesses. The alternating portions (e.g., 312(2) and 312(4)) of the removed trench dielectric portion 312 and the corresponding exposed sidewall portions comprise substantially the same thickness. Thus, two immediately adjacent recesses 330 (e.g., 330(2) and 330(3)) include different magnitudes of depth D, thickness T, or different magnitudes of depth D and thickness T (fig. 1-1). The two alternating recesses 330 (e.g., 330(2) and 330(4)) include a depth D and a thickness T of substantially the same size.

Each of the recesses 330(1), (330), (2), (330), (3), (330), (4) is covered by at least one covering layer 340 (here, 340), (1), (340), (2), (340), (3), and 340 (4)). The higher recesses 330 stacked along the sloped sidewalls 316 are covered by more layers of the capping layer 340 than the lower recesses 330 stacked along the sloped sidewalls 316, because the capping layer 340 for the lower recesses 330 also covers the higher recesses 330. In other words, the capping layer 340 accumulates over the upper recesses 330. This accumulation of the capping layer 340 is achieved by a suitable inclination of the inclined sidewall 316, i.e., a suitably formed angle θ 3. For example, in one embodiment, the angle θ 3 is between about 65 degrees and about 89 degrees. For example, the covering layer 340(4) formed to cover the lowest recess 330(4) also covers the recesses 330(3), 330(2), and 330(1) higher than the recess 330 (4). Thus, the higher recesses 330 are covered by a thicker accumulation of dielectric material (i.e., more layers of capping layer 340) than the lower recesses 330.

If the angle θ 3 is greater than about 89 degrees, the capping layer 340 may not effectively accumulate on the upper recesses 330.

In one embodiment, the thickness of the dielectric portion (e.g., 312(2)) of the removed trench 312 and the size of the corresponding recess 330(2) are controlled by the epitaxial layer formed in the space of the dielectric portion 312(2) of the removed trench 312. Different materials and/or crystallographic orientations of the epitaxial layer may include different angular crystal plane edge portions. The conditions (e.g., temperature) of the epitaxial process may also affect the edge portions of the angled crystal planes of the epitaxial layers formed. All of these factors are considered independently or collectively in determining the size and/or profile of the recess 330.

Referring back to fig. 2, and also to fig. 3I, in an exemplary operation 240, the lowermost recess 330 (here 330(4)) is exposed by etching. The etching conditions are controlled to remove the capping layer 340(4) covering the lowest recess 330(4), while the capping layer 340(3) covering the adjacent higher recess 330(3) remains and the adjacent higher recess 330(3) remains covered. Even higher recesses 330(2) and 330(1) remain covered due to the accumulation of the capping layer 340.

In example operation 245, an epitaxial layer 350(1) is formed in trench 310, wherein a crystallographic edge 352(1) of epitaxial layer 350(1) is aligned into angular recess 330 (4). Epitaxial layer 350(1) includes a first semiconductor material, such as silicon. In one embodiment, the epitaxial process includes etching features that prevent silicon growth through the interface of the cap layer 340. Therefore, epitaxial layer 350(1) does not extend beyond upper edge 332(4) of recess 330 (4). Furthermore, the thickness of epitaxial layer 350(1) may be monitored in-situ during the epitaxial process, for example, via an oscillating quartz resonator. This monitoring provides substantially real-time feedback of thickness control of the thin film deposition component of the epitaxial process. Thus, accurate control of the thickness of epitaxial layer 350(1) is enhanced by dynamically optimizing epitaxial parameters via real-time monitoring and feedback methods during the deposition process.

Because edge 352(1) of epitaxial layer 350(1) is aligned in angled recess 330(4), upper surface 354(1) of epitaxial layer 350(1) is substantially planar. The epitaxial layer thus extends substantially throughout the space defined by the upper edge 332(4) of the recess 330 (4). By aligning into angled recess 330(4), the crystallographic edge 352(1) of epitaxial layer 350(1) need not fit perfectly into recess 330 (4). In one embodiment, even if edge 352(1) of epitaxial layer 350(1) does not perfectly fit into recess 330(4), the planarity of surface 354(1) may be improved. In another embodiment, as shown in FIG. 3I-1, the profile of recess 330(4) is controlled such that edge portion 352(1) of the crystallographic plane of epitaxial layer 350(1) is completely contained within recess 330 (4). For example, the etch rate difference between crystal plane 332 and crystal plane 334 of recess 330 may be controlled/selected to achieve a desired profile of recess 330 (fig. 3F-1).

Referring to fig. 3J, operations 240 and 245 are repeated to form epitaxial layers 350(2), 350(3), 350(4) as a stack of epitaxial layers within trench 310. The edge portions 352(2), 352(3), 352(4) of each of the epitaxial layers 350(2), 350(3), 350(4) are aligned in the corresponding recesses 330(3), 330(2), 330(1), respectively.

Closely adjacent epitaxial layers 350 (e.g., 350(2) and 350(3)) comprise different semiconductor materials. Alternating epitaxial layers 350 (e.g., 350(2) and 350(4)) comprise the same semiconductor material. In one embodiment, epitaxial layers 350(1) and 350(3) are silicon or other suitable semiconductor material. Epitaxial layers 350(2) and 350(4) are silicon germanium or other suitable semiconductor materials. Epitaxial layers 350(1), 350(3) of silicon include the same thickness t 2. Epitaxial layers 350(2), 350(4) of sige include the same thickness t 3. Depending on the device and/or circuit design, t2 may or may not be equal to t 3. As provided by the manufacturing process, t3 is substantially equal to t1 (fig. 3E). Thus, in essence, the stack of epitaxial layers 350 includes silicon layers 350(1), 350(3) and silicon germanium layers 350(2), 350(4) stacked in alternating sequence. Thickness t2 of silicon epitaxial layers 350(1), 350(3) and thickness t1 of silicon germanium epitaxial layers 350(2), 350(4) are determined based on the specific device design. In the GAA process, one type of epitaxial layers (silicon epitaxial layers 350(1), 350(3) or silicon germanium epitaxial layers 350(2), 350(4)) will be used to form the channel layers of the transistor devices, and the other type will be used as sacrificial layers.

Because edge portions of each of epitaxial layers 350 are aligned into respective recesses 330, each epitaxial layer 350 is substantially planar and extends laterally all the way within trench 312.

Referring back to fig. 2, and also to fig. 3K, in an example operation 250, the sacrificial substrate region 310 is at least partially removed. The bank structure 314 of the sacrificial substrate 310 may be separately removed from the portion of the sacrificial substrate 310 positioned below the trench 312. In one embodiment, a dielectric material may be formed in the trench 312. In another embodiment, the sacrificial substrate 310 material may remain under the trenches 312. Specifically, the capping layer 320 and the pad oxide layer 322 are sequentially removed to expose the sacrificial substrate 310. Sacrificial substrate 310 (e.g., germanium) is selectively etched while epitaxial layer 350 remains. A dielectric layer 360 is then deposited within the space vacated by the removal of the sacrificial substrate 310. Accordingly, dielectric layer 360 also includes a recess that interfaces with the crystal plane edge portion 352 of epitaxial layer 350.

In other embodiments, the "sacrificial" substrate region 310 is not removed. Furthermore, as described herein, in other embodiments, sacrificial substrate region 310 is not formed and epitaxial layer 350 is formed directly within N-well 104 using the processes described herein.

Fig. 3K illustrates a planarization process (e.g., CMP) performed such that upper surface 362 of dielectric layer 360 is coplanar with (e.g., substantially at the same level as) topmost epitaxial layer 350 (here, sige 350(4)), without limiting the scope of the present disclosure.

Referring back to fig. 2, also referring to fig. 3L, in an example operation 255, the stack of epitaxial layers 350 is patterned to form a stack of nanowire strips 370 comprising silicon nanowire strips 370(1), 370(3), and silicon germanium nanowire strips 370(2), 370 (4). The crystal plane edge portions 350R of the epitaxial layer 350 may remain within the recess of the dielectric layer 360 or may be removed.

A similar process is also performed in P-well 106 to create nanowire stack 470. As shown in fig. 3M, the nanowire stack 470 includes silicon germanium nanowire strips 470(1), 470(3) and silicon nanowire strips 470(2), 470(4) stacked in an alternating sequence, wherein the silicon nanowire strip 470(4) is stacked as the topmost strip. Since nanowire stack 470 is fabricated separately from nanowire stack 370, the sequential order of nanowire strips 470 may be different from the sequential order of nanowire strips 370. For example, as illustratively shown in fig. 3M, the topmost nanowire strip in nanowire stack 370 is silicon germanium nanowire strip 370(4), and the topmost nanowire strip in nanowire stack 470 is silicon nanowire strip 470 (4). Further, the thickness of the silicon nanowire strips 470(4), 470(2) may be different from the thickness of the silicon nanowire strips 370(3), 370 (1). Similarly, the thickness of silicon germanium nanowire strips 470(3), 470(1) may be different from the thickness of silicon germanium nanowire strips 370(4), 370 (2). Within each nanowire stack 370, 470, the silicon nanowire strip may comprise a different thickness than the silicon germanium nanowire strip. Furthermore, in one embodiment, the silicon germanium nanowire strips 370(4), 370(2) are not aligned with the silicon germanium nanowires 470(3), 470(1), and are not aligned with the silicon nanowire strips 470(4), 470 (2). For example, the centerlines of such nanowire strips may be offset from one another. The silicon nanowire strips 370(3), 370(1) are not aligned with the silicon nanowire strips 470(4), 470(2), and are not aligned with the silicon germanium nanowire strips 470(3), and 470 (1). All thicknesses of the nanowire strips may be optimized separately for nanowire stack 370 and nanowire stack 470, depending on the device to be fabricated therefrom.

In one example, the nanowire stack 370 is used to fabricate a PMOS device with channel regions fabricated from silicon germanium nanowire strips 370(2), 370 (4). The silicon nanowire strips 370(1), (370), (3) are removed as sacrificial strips and replaced with gate structures. Thus, in the resulting PMOS transistor, the gate structure surrounds the surface of the silicon germanium nanowire strip.

The nanowire stack 470 is used to fabricate NMOS devices having channel regions fabricated from silicon nanowire strips 470(2), 470 (4). The silicon germanium nanowire strips 470(1), 470(3) are removed as sacrificial strips and replaced with gate structures. Thus, in the resulting NMOS transistor, the gate structure surrounds the surface of the silicon nanowire strip.

Fig. 1 illustrates example CMOS components of an integrated circuit fabricated from a wafer 300. The example process 200 of fig. 2 may be used to fabricate the example structure 100 of fig. 1 and/or other structures/devices, each of which is included in the present disclosure. Note that in fig. 1, the sequence of nanowire strips 160 differs from the sequential order of nanowire stacks 470 shown in fig. 3M. The sequential order, materials, and thicknesses or other parameters of the nanowire strips in the locally formed nanowire stacks 370, 470 may be customized and optimized for each sacrificial substrate region 310, 410 and based on design requirements, as described herein. This flexibility facilitates improving the device performance of PMOS and NMOS devices separately and independently. In addition, the insulating layers 360, 460 are integrated when forming the nanowire stacks 370, 470, which simplifies the integration of CMOS processes into high voltage analog processes, such as bipolar-CMOS-dmos (bcd) processes.

In addition, the disclosed technique for forming an epitaxial layer solves the problems generated by the edge portions of the angular crystal plane of the epitaxial layer. The resulting epitaxial layer is substantially planar and extends in the entire trench of the local epitaxial growth region (i.e., the sacrificial substrate region). Thereby, the product quality can be improved and the size of the space between the devices can be further reduced.

The disclosure may be further understood with reference to the following description of embodiments:

in one embodiment, an integrated circuit includes a substrate, a first device, a second device, and a first insulating structure between the first device and the second device. The first device includes a first stack of isolated nanowire structures of a first semiconductor material over a substrate, and a first gate structure surrounding the first stack of isolated nanowire structures. The second device includes a second stack of isolated nanowire structures of a second semiconductor material over the substrate, and a second gate structure surrounding the second stack of isolated nanowire structures.

In one embodiment, the integrated circuit further comprises a second insulating structure that is one or more of laterally between the first device and the second device, or longitudinally between the substrate and the second device.

In one embodiment, the first stack of split nanowire structures includes a first split nanowire structure and a second split nanowire structure vertically adjacent to the first split nanowire structure, the second split nanowire structure is vertically separated from the first split nanowire structure by a portion of the first gate structure, and a thickness of the portion of the first gate structure is different in a vertical direction from a thickness of a third split nanowire structure of the second stack of split nanowire structures.

In one embodiment, the first semiconductor material is silicon and the second semiconductor material is silicon germanium.

In one embodiment, the first insulating structure is embedded in a first doped region of the substrate.

In one embodiment, the second insulating structure is embedded in a second doped region of the substrate.

In another embodiment, a method includes forming a second substrate region within a first substrate region. A trench is formed in the second substrate region, and a bank structure of the second substrate region is formed around the trench. The bank structure includes sloped sidewall portions and upper corners. The trench is filled with a different material than the second substrate region. The first sidewall portion of the sloped sidewall is exposed. A first recess is formed in the exposed first sidewall portion. A first dielectric capping layer is formed overlying the first recess. Exposing a second sidewall portion of the sloped sidewall. The second sidewall portion is below the first sidewall portion. A second depression is formed in the exposed second sidewall portion below the first angular depression. A second dielectric capping layer is formed overlying the second recess and extending over the first dielectric capping layer. The second recess is exposed while the first recess remains covered. A first semiconductor epitaxial layer is formed in the trench, wherein a crystallographic plane edge portion of the first semiconductor epitaxial layer is aligned into the second recess. Exposing the first angular depression. A second semiconductor epitaxial layer is formed in the trench above the first semiconductor epitaxial layer, wherein a crystallographic edge portion of the second semiconductor epitaxial layer is aligned into the first recess.

In one embodiment, the method further comprises forming a dielectric capping layer covering the upper corners of the trench.

In one embodiment, the method further comprises forming an oxide layer between the dielectric capping layer and an upper surface of the bank.

In one embodiment, the first sidewall portion and the second sidewall portion include a different thickness in a vertical direction.

In one embodiment, the inclined sidewall defines a space that is substantially a convex isosceles trapezoid in a cross-sectional view.

In one embodiment, the method further comprises: forming an aperture by removing a remaining portion of the bank structure; and forming an insulating body by filling the aperture with a dielectric material.

In one embodiment, filling the aperture with a dielectric material comprises: a planarization process is performed to make the insulating body and the second semiconductor epitaxial layer substantially at a same level.

In one embodiment, the method further comprises: forming a vertical stack of a first semiconductor strip and a second semiconductor strip by patterning the first semiconductor epitaxial layer and the second semiconductor epitaxial layer; removing one of the first semiconductor strip or the second semiconductor strip; and forming a gate structure surrounding at least three surfaces of the other of the first semiconductor strip or the second semiconductor strip.

In one embodiment, the method further comprises: removing a portion of the second substrate region under the trench; and forming an insulating body below the trench.

In an embodiment, a trench is formed extending down through the second substrate region and reaching the first substrate region.

In one embodiment, the first substrate region is a doped substrate region.

In a further embodiment, a method includes receiving a wafer including a substrate having a P-well and an N-well. The first substrate body is formed in the P-well, and the second substrate body is formed in the N-well. The first trench is formed in the first substrate body, and the second trench is formed in the second substrate body. The first trench and the second trench have sloped sidewalls. A first recess is formed on the sloped sidewall of the first trench. The second recess is formed on the sloped sidewall of the second trench. The second recess has a different thickness than the first recess. A first epitaxial layer of a first semiconductor material is formed in the first trench with a crystalline surface edge portion of the first semiconductor epitaxial layer extending into the first recess. A second epitaxial layer of the first semiconductor material is formed in the second trench, wherein a crystallographic plane edge portion of the second semiconductor epitaxial layer extends into the second recess. The second epitaxial layer has a different thickness than the first epitaxial layer.

In one embodiment, the first and second epitaxial layers are not aligned with each other in a lateral direction.

In one embodiment, the method further comprises: forming an aperture by removing at least one of the first substrate body or the second substrate body; and forming an insulating body by filling a dielectric material into the void.

The foregoing outlines features of several embodiments so that those skilled in the art may better understand the aspects of the present disclosure. Those skilled in the art should appreciate that the present disclosure may be readily utilized as a basis for designing or modifying other processes and structures for carrying out the same purposes and/or achieving the same advantages of the embodiments introduced herein. Those skilled in the art should also realize that such equivalent constructions do not depart from the spirit and scope of the present disclosure, and that they may make various changes, substitutions, and alterations herein without departing from the spirit and scope of the present disclosure.