阅读说明：本技术 三维存储器件及其制作方法 (Three-dimensional memory device and manufacturing method thereof ) 是由 吴林春 于 2020-04-30 设计创作，主要内容包括：提供了三维(3D)NAND存储器件和方法。在一个方面,一种3D NAND存储器件包括：衬底、处于衬底之上的层堆叠体、第一外延层、第二外延层、第一阵列公共源极(ACS)和第二ACS。层堆叠体包括交替堆叠设置的第一堆叠层和第二堆叠层。第一外延层被沉积在穿过层堆叠体延伸的沟道层的侧面部分上。第二外延层被沉积在衬底上。第一ACS以及层堆叠体的部分处于第二ACS之间。(Three-dimensional (3D) NAND memory devices and methods are provided. In one aspect, a 3D NAND memory device includes: a substrate, a layer stack over the substrate, a first epitaxial layer, a second epitaxial layer, a first Array Common Source (ACS), and a second ACS. The layer stack includes first stacked layers and second stacked layers alternately stacked. A first epitaxial layer is deposited on side portions of the channel layer extending through the layer stack. A second epitaxial layer is deposited on the substrate. The first ACS and the portion of the layer stack are between the second ACS.)

1. A three-dimensional (3D) memory device, comprising:

a substrate;

a layer stack over the substrate, the layer stack including a plurality of first stacked layers and a plurality of second stacked layers alternately arranged in a stack, and a plurality of memory blocks;

a first epitaxial layer on a side portion of a channel layer extending through the layer stack;

a second epitaxial layer on the substrate below the layer stack; and

a plurality of first Array Common Source (ACS) and a plurality of second ACS configured to extend through the layer stack and electrically connected with the second epitaxial layer, wherein the plurality of second ACS extend continuously along a first direction, thereby dividing the layer stack into the plurality of storage blocks, the plurality of first ACS are located within one storage block and extend discontinuously in the first direction.

2. The device of claim 1, wherein the first epitaxial layer is contiguous with the second epitaxial layer, and a portion of the layer stack and the plurality of first ACS are between the plurality of second ACS.

3. The device of claim 1, wherein the plurality of first ACS partially overlaps an area of a Top Select Gate (TSG) cut.

4. The device of claim 1, further comprising:

a functional layer deposited on sidewalls and a bottom surface of a channel hole extending through the layer stack; and

a dielectric material filling the channel hole,

wherein the channel layer is deposited adjacent to the functional layer and the functional layer comprises a barrier layer, a charge trapping layer, and/or a tunneling insulation layer.

5. The device of claim 1, wherein the plurality of first ACS and the plurality of second ACS each comprise:

an electrically isolating layer deposited on sidewalls of the gate Gap (GLS); and

a conductive material deposited on the electrically isolating layer in the GLS.

6. The device of claim 1, wherein the plurality of first stack layers comprise a conductive material.

7. The device of claim 1, wherein the plurality of first ACS each form a cylindrical shape or a columnar shape.

8. The device of claim 1, wherein the plurality of first ACS comprise different shapes.

9. The device of claim 1, wherein the plurality of first ACS are spaced apart by a predetermined distance.

10. The device of claim 1, wherein the plurality of first ACS are spaced apart by distances of different values.

11. The device of claim 1, wherein the plurality of first ACS forms a two-dimensional pattern.

12. The device of claim 1, wherein the plurality of first ACS forms a plurality of rows each extending in a direction parallel to the plurality of second ACS.

13. A method for fabricating a three-dimensional (3D) memory device, comprising:

forming a layer stack over a substrate, the layer stack comprising a plurality of first stacked layers and a plurality of second stacked layers arranged in an alternating stack;

forming a plurality of first gate slits (GLS) and a plurality of second GLS extending through the layer stack, wherein the plurality of second GLS continuously extend in a first direction to divide the layer stack into a plurality of storage blocks, the plurality of first GLS are located in one storage block and discontinuously extend in the first direction,

performing epitaxial growth by means of the first GLS and the second GLS to deposit a first epitaxial layer on a side portion of a channel layer extending through the layer stack and to deposit a second epitaxial layer on the substrate below the layer stack.

14. The method of claim 13, wherein the first epitaxial layer is contiguous with the second epitaxial layer, and a portion of the layer stack and the plurality of first GLSs are between the plurality of second GLSs.

15. The method of claim 13, further comprising:

forming a channel hole extending through the layer stack;

forming a functional layer on sidewalls and a bottom surface of the channel hole, the functional layer including a barrier layer, a charge trap layer, and/or a tunneling insulating layer; and

filling the channel hole with a dielectric material,

wherein the channel layer is deposited adjacent to the functional layer.

16. The method of claim 15, further comprising:

prior to performing the epitaxial growth, removing portions of the functional layer to expose the side portions of the channel layer to grow the first epitaxial layer on the side portions and the second epitaxial layer on the substrate.

17. The method of claim 13, further comprising:

depositing a sacrificial layer over the substrate;

depositing a spacer layer on a sidewall and a bottom surface of one of the plurality of first GLSs;

removing a portion of the spacer layer on the bottom surface by etching to expose a portion of the sacrificial layer; and

removing the sacrificial layer by etching before performing the epitaxial growth to form a cavity.

18. The method of claim 13, further comprising:

removing the capping layer on the substrate by etching before performing the epitaxial growth, thereby exposing a surface of the substrate.

19. The method of claim 13, further comprising:

filling the plurality of first GLSs with at least one conductive material to form a plurality of first array common source electrodes (ACS); and

filling the plurality of second GLSs with at least one conductive material to form a plurality of second ACSs,

wherein the first plurality of ACS and the second plurality of ACS are electrically connected to the second epitaxial layer.

20. The method of claim 13, further comprising:

portions of the plurality of first stack layers are removed by an etching process.

21. The method of claim 20, further comprising:

forming a plurality of conductor layers, wherein the plurality of conductor layers and the plurality of second stacked layers are alternately stacked.

22. A method for fabricating a three-dimensional (3D) memory device, comprising:

forming a layer stack over a substrate, the layer stack comprising a plurality of first stacked layers and a plurality of second stacked layers arranged in an alternating stack;

forming a plurality of first gate slits (GLS) and a plurality of second GLS extending through the layer stack;

performing epitaxial growth by means of the first GLS and the second GLS, thereby depositing a first epitaxial layer on a side portion of a channel layer extending through the layer stack and depositing a second epitaxial layer on the substrate below the layer stack; and

forming a plurality of first Array Common Source (ACS) and a plurality of second ACS extending through the layer stack by filling the plurality of first GLSs and the plurality of second GLSs with a conductive material, respectively, such that the plurality of first ACS and the plurality of second ACS are electrically connected with the second epitaxial layer, wherein the plurality of second ACS extend continuously in a first direction, thereby dividing the layer stack into the plurality of memory blocks, the plurality of first ACS being located in one memory block and extending discontinuously in the first direction.

23. The method of claim 22, wherein the first epitaxial layer is contiguous with the second epitaxial layer, and a portion of the layer stack and the plurality of first ACS are between the plurality of second ACS.

24. The method of claim 22, further comprising:

forming a channel hole extending through the layer stack; and

forming a functional layer on sidewalls of the channel hole, the functional layer including a barrier layer, a charge trapping layer and/or a tunneling insulating layer,

wherein the channel layer is deposited adjacent to the functional layer.

25. The method of claim 24, further comprising:

prior to performing the epitaxial growth, removing portions of the functional layer to expose the side portions of the channel layer to grow the first epitaxial layer on the side portions and the second epitaxial layer on the substrate.

26. The method of claim 22, further comprising:

depositing a sacrificial layer over the substrate; and

and removing the sacrificial layer by etching to form a cavity before the epitaxial growth is carried out.

27. The method of claim 22, further comprising:

and removing the covering layer on the substrate by etching before the epitaxial growth is carried out so as to expose the surface of the substrate.

28. The method of claim 22, further comprising:

portions of the plurality of first stack layers are removed by an etching process.

29. The method of claim 28, further comprising:

forming a plurality of conductor layers, wherein the plurality of conductor layers and the plurality of second stacked layers are alternately stacked.

30. The method of claim 22, wherein the plurality of first ACS partially overlaps an area of a Top Select Gate (TSG) cut.

Technical Field

The present application relates to the field of semiconductor technology, and in particular, to three-dimensional (3D) semiconductor memory devices and methods of fabricating the same.

Background

NAND (NAND) memory is a type of non-volatile memory that does not require power to retain stored data. The ever-increasing demand for consumer electronics, cloud computing, and big data has brought on a continuing demand for larger capacity, higher performance NAND memories. Conventional two-dimensional (2D) NAND memories are approaching their physical limits and three-dimensional (3D) NAND memories are now playing an important role. The 3D NAND memory uses multiple stacked layers in a single chip to achieve higher density, higher capacity, faster performance, lower power consumption, and better economic efficiency.

During the fabrication of the 3D NAND memory device, a gate slit (GLS) is formed to expose a sacrificial layer above a substrate. Thereafter, a cavity is etched and selective epitaxial growth of single crystal silicon and polycrystalline silicon is performed in the cavity. Since the epitaxial growth is faster near the cavity opening, a void will be left in the middle of the cavity when the opening is filled. These voids can lead to leakage of current and reliability problems.

GLS is also used to form the gate electrode in the layer stack. The sacrificial stack layers of the layer stack are etched away before the gate electrode is fabricated. However, some portions of the sacrificial layer farther from the GLS tend not to be completely etched away. Thus, some portions of the gate electrode may be only partially fabricated, which would result in failure of the NAND memory cell.

The disclosed methods and systems are directed to solving one or more of the problems set forth above, as well as other problems.

Disclosure of Invention

In one aspect of the present disclosure, a 3D NAND memory device includes: a substrate, a layer stack over the substrate, a first epitaxial layer, a second epitaxial layer, a first Array Common Source (ACS), and a second ACS. The layer stack includes a storage block and first and second stacked layers alternately stacked. A first epitaxial layer is deposited on side portions of the channel layer extending through the layer stack. A second epitaxial layer is deposited on the substrate. The first ACS and the second ACS are configured for each memory block and extend through the layer stack. The first epitaxial layer adjoins the second epitaxial layer. The first ACS and the second ACS are electrically connected to the second epitaxial layer. A portion of the layer stack and the first ACS are between the second ACS.

In another aspect of the present disclosure, a method for fabricating a 3D NAND memory device includes: forming a layer stack over a substrate; performing epitaxial growth to deposit a first epitaxial layer on side portions of the channel layer extending through the layer stack and to deposit a second epitaxial layer on the substrate; and forming a first gate slit (GLS) and a second GLS for each memory block extending through the layer stack. The layer stack includes first stacked layers and second stacked layers alternately stacked. The first epitaxial layer adjoins the second epitaxial layer. The first GLS and portions of the layer stack are between the second GLS.

In another aspect of the present disclosure, another fabrication method for a 3D NAND memory device includes: forming a layer stack over a substrate; performing epitaxial growth to deposit a first epitaxial layer on side portions of the channel layer extending through the layer stack and to deposit a second epitaxial layer on the substrate; and forming a first Array of Common Sources (ACS) and a second ACS for each memory block extending through the layer stack. The layer stack includes first stacked layers and second stacked layers alternately stacked. The first epitaxial layer adjoins the second epitaxial layer. The first ACS and the second ACS are electrically connected to the second epitaxial layer. A portion of the layer stack and the first ACS are between the second ACS.

Other aspects of the disclosure will become apparent to those skilled in the art from consideration of the specification, claims and drawings presented herein.

Drawings

FIG. 1 schematically illustrates a cross-sectional view of a three-dimensional (3D) memory device in an exemplary fabrication process according to an embodiment of the disclosure;

fig. 2 and 3 schematically illustrate top and cross-sectional views of the 3D memory device shown in fig. 1 after forming a channel hole according to an embodiment of the present disclosure;

fig. 4 and 5 schematically illustrate top and cross-sectional views of the 3D memory device shown in fig. 2 and 3 after forming a gate slit (GLS) according to an embodiment of the present disclosure;

FIGS. 6 and 7 schematically illustrate cross-sectional views of the 3D memory device shown in FIGS. 4 and 5 after deposition and subsequent selective etching of GLS spacers, in accordance with an embodiment of the present disclosure;

FIGS. 8 and 9 schematically illustrate cross-sectional views of the 3D memory device shown in FIG. 7 after performing certain etching steps, in accordance with embodiments of the present disclosure;

FIG. 10 schematically illustrates a cross-sectional view of the 3D memory device shown in FIG. 9 after selective epitaxial growth, in accordance with an embodiment of the present disclosure;

fig. 11 schematically illustrates a cross-sectional view of the 3D memory device shown in fig. 10 after forming a conductor layer according to an embodiment of the present disclosure;

FIG. 12 schematically illustrates a cross-sectional view of the 3D memory device shown in FIG. 11 after forming an Array Common Source (ACS) according to an embodiment of the present disclosure;

FIG. 13 shows a schematic flow diagram of the fabrication of a 3D memory device according to an embodiment of the present disclosure;

fig. 14 and 15 schematically illustrate top and cross-sectional views of another 3D memory device after forming a GLS according to an embodiment of the present disclosure;

FIG. 16 schematically illustrates a top view of another 3D memory device, according to an embodiment of the present disclosure;

fig. 17 and 18 schematically illustrate top and cross-sectional views of another 3D memory device after forming a GLS according to an embodiment of the present disclosure; and is

Fig. 19 and 20 schematically illustrate cross-sectional and top views of the 3D memory device shown in fig. 17 and 18 after several fabrication steps, in accordance with an embodiment of the present disclosure;

fig. 21 and 22 schematically illustrate top and cross-sectional views of the 3D memory device shown in fig. 17 and 18 with additional features according to embodiments of the present disclosure.

Detailed Description

Technical solutions in embodiments of the present disclosure will be described below with reference to the accompanying drawings. Wherever possible, the same reference numbers will be used throughout the drawings to refer to the same or like parts. It is to be understood that the described embodiments are merely some, and not all, embodiments of the disclosure. Features from various embodiments may be interchanged and/or combined. Other embodiments obtained based on the embodiments of the disclosure by those skilled in the art without inventive effort should fall within the scope of the disclosure.

Fig. 1-12 schematically illustrate a fabrication process of an exemplary 3D memory device 100 according to an embodiment of the present disclosure. In fig. 1-12, the cross-sectional view is in the Y-Z plane and the top view is in the X-Y plane. As shown in fig. 1, the 3D memory device 100 includes a substrate 110. In some embodiments, substrate 110 may comprise a single crystal silicon layer. In some other embodiments, the substrate 110 may comprise other semiconductor materials, such as germanium (Ge), silicon germanium (SiGe), silicon carbide (SiC), silicon-on-insulator (SOI), germanium-on-insulator (GOI), polysilicon, III-V compounds such as gallium arsenide (GaAs) or indium phosphide (InP), and the like. In some other embodiments, the substrate 110 may comprise a non-conductive material, such as glass, a plastic material, or a ceramic material. In the following description, the substrate 110 includes an undoped or lightly doped single crystal silicon layer, as an example. In some other embodiments, the substrate 110 may be doped differently with p-type or n-type dopants. When the substrate 110 comprises a glass, plastic, or ceramic material, the substrate 110 may further comprise a thin layer of polysilicon deposited on the glass, plastic, or ceramic material, such that the substrate 110 may be processed like a polysilicon substrate.

As shown in fig. 1, a capping layer 120 may be deposited over the substrate 110. The capping layer 120 is a sacrificial layer and may include a single layer or a composite layer having a plurality of layers. For example, layer 120 may include one or more of a silicon oxide layer and a silicon nitride layer. Layer 120 may be deposited by Chemical Vapor Deposition (CVD), Physical Vapor Deposition (PVD), Atomic Layer Deposition (ALD), or a combination of two or more of these methods. In some other embodiments, layer 120 may comprise other materials, for example, alumina.

Over the capping layer 120, a sacrificial layer 130 may be deposited. The sacrificial layer 130 may include a dielectric material, a semiconductor material, or a conductive material. For example, layer 130 may be single crystal silicon or polycrystalline silicon, which may be deposited by a CVD and/or PVD process. In the description that follows, an exemplary material for layer 130 is polysilicon. After the polysilicon layer 130 is formed, the layer stack 140 may be deposited. Layer stack 140 includes a plurality of pairs of stacked layers 141 and 142, i.e., layers 141 and 142 are alternately stacked. For example, the layer stack may include 64 pairs, 128 pairs, or more than 128 pairs of layers 141 and 142.

In some embodiments, layers 141 and 142 may each comprise a first dielectric material and a second dielectric material different from the first dielectric material. Alternating layers 141 and 142 may be deposited via CVD, PVD, ALD, or a combination of two or more of these processes. In the discussion that follows, exemplary materials for layers 141 and 142 are silicon oxide and silicon nitride, respectively. The silicon oxide layer 141 may be configured as an isolation layer, and the silicon nitride layer 142 may be configured as a sacrificial layer. The sacrificial stack layer 142 will be etched away during the fabrication process and replaced by a conductor layer. In some other embodiments, different materials may be used to form alternating stacked layers 141 and 142. For example, layers 141 and 142 may comprise dielectric materials other than silicon oxide and/or silicon nitride. Furthermore, in some other embodiments, layers 141 and 142 may include dielectric and conductive layers. The conductive layer may comprise, for example, tungsten (W), cobalt (Co), copper (Cu), aluminum (Al), doped silicon, or silicide. In the discussion that follows, layers 141 and 142 comprise silicon oxide and silicon nitride, respectively, as previously described. In addition, the silicon nitride layer 142 is configured as a sacrificial layer to be etched away in the fabrication process.

Fig. 2 and 3 schematically illustrate top and cross-sectional views of the 3D memory device 100 after forming and then filling the channel holes 150, according to an embodiment of the present disclosure. The number, size, and arrangement of the channel holes 150 shown in the figures are exemplary only and are used to describe the structure and fabrication method of the device 100. The channel holes 150 are configured to extend in the Z direction and form an array of predetermined patterns in the X-Y plane. The cross-sectional view shown in fig. 3 is taken along line AA' of fig. 2. Thus, FIG. 3 shows only some of the trench holes 150 of FIG. 2 in cross-section in the Y-Z plane.

The channel hole 150 may be formed, for example, by a dry etching process or a combination of a dry etching process and a wet etching process. Other fabrication processes, such as patterning processes involving photolithography, cleaning, and/or Chemical Mechanical Polishing (CMP), may also be performed, and a detailed description of these processes is omitted for simplicity. The channel hole 150 may have a cylindrical or columnar shape extending through the layer stack 140 and the layers 130 and 120 and partially penetrating the substrate 110. After forming the channel hole 150, a functional layer 151 may be deposited on the sidewalls and bottom of the channel hole. The functional layer 151 may include a blocking layer 152 on sidewalls and a bottom of the channel hole to block outflow of charges, a charge trap layer 153 on a surface of the blocking layer 152 to store charges during operation of the 3D memory device 100, and a tunneling insulation layer 154 on a surface of the charge trap layer 153. The barrier layer 152 may include one or more layers, which may include one or more materials. The material of the barrier layer 152 may include silicon oxide, silicon nitride, silicon oxynitride, a high-k dielectric material (e.g., aluminum oxide or hafnium oxide), another wide bandgap material, and the like. The charge trapping layer 153 may include one or more layers, which may include one or more materials. The material of the charge trapping layer 153 may include polysilicon, silicon nitride, silicon oxynitride, nanocrystalline silicon, other wide band gap materials, and the like. The tunneling insulating layer 154 may include one or more layers, which may include one or more materials. The material of the tunneling insulating layer 154 may include silicon oxide, silicon nitride, silicon oxynitride, a high-k dielectric material (e.g., aluminum oxide or hafnium oxide), another wide bandgap material, and the like.

In some embodiments, the functional layer 151 may include an oxide-nitride-oxide (ONO) structure. However, in some other embodiments, the functional layer 151 may have a structure other than the ONO configuration. In the following description, an ONO structure is used. For example, the functional layer 151 may include a silicon oxide layer, a silicon nitride layer, and another silicon oxide layer.

As shown in fig. 3, a silicon oxide layer may be deposited on the sidewalls of the trench hole 150 as a barrier layer 152. A silicon nitride layer may be deposited on the barrier layer 152 as a charge trapping layer 153. Another silicon oxide layer may be deposited on the charge trapping layer 153 as a tunneling insulating layer 154. A polysilicon layer may be deposited on the tunnel insulation layer 154 as a channel layer 155, the channel layer 155 also being referred to as a "semiconductor channel". In some other embodiments, the channel layer 155 (semiconductor channel) may include amorphous silicon. Similar to the channel holes, channel layers 155 (semiconductor channels) also extend through the layer stack 140 and into the substrate 110, as shown in fig. 3, with a portion of each functional layer 151 disposed between a portion of one of the stack layers 141 and 142 and a portion of one of the channel layers 155. The barrier layer 152, charge-trapping layer 153, tunneling insulation layer 154, and channel layer 155 may be deposited by, for example, CVD, PVD, ALD, or a combination of two or more of these processes. The channel hole 150 may be filled with an oxide material 156 after the channel layer 155 is formed.

Fig. 4 and 5 schematically illustrate a top view and a cross-sectional view of the 3D memory device 100 after forming a gate slit (GLS) according to an embodiment of the present disclosure. The cross-sectional view shown in fig. 5 is taken along line BB' of fig. 4. The 3D memory device 100 may have a large number of NAND memory cells configured in the layer stack 140. The layer stack 140 may be divided into a plurality of storage blocks. In some embodiments, NAND memory cells belonging to a memory block may be reset together in a block erase operation. As shown in fig. 4, a memory block region 101 corresponding to a memory block may include a channel hole region 102, and the channel hole region 102 may include a first gate slit (GLS) region 160. The memory block area 101 may be separated from other memory blocks (not shown) by a pair of GLSs 170 representing the second GLS area. As shown in fig. 4, the memory block area 101 is arranged between the pair of GLSs 170 in the Y direction.

The first GLS region 160 may include a plurality of GLSs 161, the plurality of GLSs 161 being in a middle portion of the trench hole region 102 between the GLSs 170. In some embodiments, the GLS 161 may have the same shape as the channel hole 150 and have a size similar to the size of the channel hole 150. In some embodiments, the GLS 161 may have a shape different from the shape of the channel hole 150 and/or a size different from the size of the channel hole 150. For example, the GLS 161 may have a circular shape, a square shape, a diamond shape, an oval shape, and the like. The first GLS region 160 may include a certain number of GLS 161, and in some embodiments, the first GLS region 160 may extend in the X-direction. In some other embodiments, the first GLS region 160 may extend in both the X-direction and the Y-direction. The GLS 161 in the first GLS region 160 are configured to be discontinuous and spaced apart from each other in the X direction. That is, adjacent GLSs 161 are spaced apart from each other and an interval between two adjacent GLSs 161 may be at least greater than a predetermined value. Since the GLS 170 continuously extends in the X direction, the arrangement of the GLS 161 may be regarded as extending in a direction parallel to the GLS 170 and spaced apart. In some embodiments, the interval between two adjacent GLSs 161 may be a fixed value. In some other embodiments, the interval between two adjacent GLSs 161 may have different values. The interval between adjacent GLSs 161 is configured such that the gate electrode of each NAND memory cell in the memory block region 101 is electrically connected. As such, the first GLS region 160 or the GLS 161 splits the channel hole region 102 into two electrically connected portions, and the function of the memory block region 101 is not affected by the GLS 161.

GLS 170 is continuous in the X direction, as compared to discontinuous GLS 161. As shown in fig. 4, the pair of GLSs 170 are parallel to each other and each extend continuously in the X-direction from the left side to the right side of the trench hole region 102. The pair of GLSs 170 representing the second GLS area may be regarded as two boundary areas of the memory block area 101. Thus, pairs of GLSs 170 may divide the NAND memory cells of device 100 into a plurality of memory block areas (not shown).

GLS 161 and GLS 170 may be formed by, for example, a dry etching process or a combination of a dry etching process and a wet etching process. As shown in fig. 5, both GLS 161 and GLS 170 extend in the Z-direction through the layer stack 140 and reach or partially penetrate the polysilicon layer 130. As such, portions of the polysilicon layer 130 are exposed at the bottom of the GLS 161 and the GLS 170. Because of GLS 161, more of the layer 130 or a larger total area is exposed than in the absence of GLS 161. Thus, it becomes faster and easier to etch away the layer 130 to form the cavity, and it becomes faster and easier to grow the epitaxial layer in the cavity. In addition, a greater portion or total area of the stack layer 142 is exposed because of the GLS 161. Similarly, it becomes faster and easier to etch away layer 142 to form the cavity, and the cavity can be filled with conductive material faster and more easily.

Fig. 6 and 7 schematically illustrate cross-sectional views of a 3D memory device after deposition and subsequent selective etching of GLS spacers, according to an embodiment of the disclosure. As shown in fig. 6 and 7, the GLS spacer includes layers 171, 172, 173, and 174 that may be deposited sequentially by CVD, PVD, ALD, or a combination of two or more of these processes. Layers 171 and 173 may comprise, for example, silicon nitride, and layers 172 and 174 may comprise, for example, silicon oxide. After depositing the GLS spacer, a selective etch is performed such that portions of the spacer at the bottom of the GLS 161 and the GLS 170 are removed by a dry etch or a combination of a dry etch and a wet etch. As such, the polysilicon layer 130 is partially exposed at the bottom of the GLS 161 and the GLS 170, as shown in fig. 7.

Fig. 8-12 each schematically illustrate a cross-sectional view of the 3D memory device 100 after performing one or more etching steps, in accordance with an embodiment of the present disclosure. A first selective etching process (e.g., a selective wet etching process) is performed to remove the polysilicon material of the polysilicon layer 130. As shown in fig. 8, the removal of the polysilicon material creates a cavity 180, exposing the capping layer 120 and a bottom portion of the silicon oxide layer (i.e., the barrier layer 152) of the functional layer 151 described above formed in the channel hole 150.

After the polysilicon layer 130 is etched, a second selective etching process (e.g., a selective wet etching process) is performed to remove a portion of the silicon oxide layer of the functional layer 151 exposed in the cavity 180. As a result, a portion of the silicon nitride layer (i.e., the charge trap layer 153) of the functional layer 151 is exposed. Since layer 174 of the GLS spacer is silicon oxide, layer 174 is also removed in the second selective etch process. Thus, the silicon nitride layer 173 is exposed.

After exposing portions of silicon nitride layer 153 and layer 173, a third selective etch process (e.g., a selective wet etch process) is performed to remove the exposed silicon nitride material, including the exposed portions of silicon nitride charge trapping layer 153 and layer 173. The removal of the silicon nitride material exposes portions of the silicon oxide layer (i.e., the tunneling insulating layer 154) of the functional layer 151 and the silicon oxide layer 172 of the GLS spacer in the cavity 180.

Thereafter, a fourth selective etching process (e.g., a selective wet etching process) is performed to remove the exposed portion of the silicon oxide tunneling insulating layer 154. In some embodiments, silicon oxide layer 172 may be configured to be substantially thicker than layer 154. As such, only portions of layer 172 may be removed after layer 154 is etched away in the fourth selective etch. The remainder of layer 172 may form layer 1721. The removal of tunnel insulation layer 154 exposes portions of the polysilicon layer (i.e., channel layer 155) in cavity 180.

In some embodiments, the capping layer 120 may be silicon oxide. Thereafter, layer 120 may be removed while etching away the bottom portion of functional layer 151. In some other embodiments, the capping layer 120 may comprise a material other than silicon oxide, or may be a composite layer. Thereafter, the layer 120 may be removed by a fifth selective etch process.

The removal of the cap layer 120 creates a cavity 181 and exposes the top surface of the substrate 110 at the bottom of the cavity 181, as shown in fig. 9. Cavity 181 is larger than cavity 180 and has an opening 182 and an opening 183. The opening 182 and the opening 183 correspond to the GLS 161 and the GLS 170, respectively. If the opening 182 were not present, the cavity 181 would have only two openings, namely the openings 183 on the left and right sides. Thereafter, portions of the substrate in the middle and exposed portions of the channel layer 155 (i.e., sidewalls of the cavity 181) may be spaced apart from the openings 183 by a distance of about half of a distance between the openings 183. The further away from opening 183, the slower the selective epitaxial growth of silicon. Thereafter, the silicon layer may grow faster on some portions of the substrate 110 opposite the openings 183 than on some other portions of the substrate and some sidewalls further from the openings 183. Thereafter, silicon deposited on substrate 110 may have accessed and sealed openings 183 before the voids near the middle region between openings 183 are filled. That is, voids may be formed near some of the channel holes 150 (particularly those in the middle regions between the openings 183). Openings 182 are created near the middle region between openings 183 because of GLS 161. As such, the silicon layer may grow faster near the intermediate region, and thus the silicon layer may grow more uniformly in the cavity 181 and voids may be avoided.

As shown in fig. 9, after the etching process described above, a portion of the polysilicon channel layer 155 near the bottom of the channel hole 150 and the substrate 110 are exposed. Layer 1721 is also in an exposed state. Layer 1721 and layer 171 may be used to protect stack of layers 140 before stack of layers 142 is etched and replaced.

Fig. 10 schematically illustrates a cross-sectional view of a 3D memory device 100 after selective epitaxial growth, according to an embodiment of the present disclosure. Selective epitaxial growth is performed to deposit a silicon layer 184 in the cavity 181. Since the substrate 110 is single crystalline silicon and the exposed portion (sidewall) of the channel layer 155 is polycrystalline silicon, a single crystalline silicon layer is grown on the substrate and a polycrystalline silicon layer is grown on the sidewall of the cavity 181. The single crystal silicon layer and the polysilicon layer are grown simultaneously and abut or merge with each other to form layer 184 as shown in fig. 10. In some embodiments, the layer 182 may be doped with a p-type dopant.

Because the GLS 161 provide openings 182 in the intermediate regions between the GLS 170, the top surface of the substrate 110 and the sidewalls of the cavity 181 (including those in the intermediate regions) are within a certain distance relative to at least one of the openings 182 or the openings 183. Thus, the cavity 181 may be filled with the layer 184 without leaving a void. As previously described, if the GLS 161 is not present, the silicon layer may grow slower in the middle region between the openings 183 and thus may form pores over portions of the substrate 110. These voids can lead to leakage of current as well as functional and reliability issues.

Fig. 11 schematically shows a cross-sectional view of the 3D memory device 100 after performing additional fabrication steps, according to an embodiment of the present disclosure. After the selective epitaxial growth, an etching process (e.g., a selective wet etching process) may be performed to remove the silicon oxide layer 1721 and the silicon nitride layer 171. Since the layer 142 of the layer stack 140 is also a silicon nitride layer, the silicon nitride layer 142 is also removed during this etching process, leaving cavities between the silicon oxide layers 141. Thereafter, a conductive material (e.g., W) is grown to fill the cavity left by the removal of layer 142, thereby forming a conductor layer 143 between silicon oxide layers 141. That is, the conductor layer 143 replaces the dielectric layer 142, and the layer stack 140 now comprises alternating dielectric layers 141 and conductor layers 143, as shown in fig. 11. The conductor layers 143 may be parallel to the substrate 110, and a portion of each functional layer 151 in the channel hole 150 is between a portion of one of the conductor layers 143 and a portion of the channel layer 155 in the channel hole 150. The conductive material may be deposited by CVD, PVD, ALD, or a combination of two or more of these processes. In some embodiments, another metal such as Co, Cu, or Al may be used as a conductive material for forming the conductor layer 143.

Each conductor layer 143 is configured to electrically connect one or more rows of NAND memory cells in the Y direction or in the X-Y plane, and each conductor layer 143 is configured as a word line of the 3D memory device 100. The channel layer 155 formed in the channel hole 150 is configured to electrically connect a column or a string of NAND memory cells in the Z direction, and the channel layer 155 is configured as a bit line of the 3D memory device 100. As such, the portion of functional layer 151 in channel hole 150 in the X-Y plane is disposed between conductor layer 143 and channel layer 155, i.e., between a word line and a bit line, as part of a NAND memory cell. A portion of the conductor layer 143 near the portion of the channel hole 150 functions as a control gate or a gate electrode of the NAND memory cell. The 3D memory device 100 as shown in fig. 11 may be regarded as a 2D array including strings of NAND cells (such strings are also referred to as "NAND strings"). Each NAND string includes a plurality of NAND cells and extends vertically toward the substrate 110. The NAND strings form a 3D arrangement of NAND memory cells.

The GLS 161 plays an important role in etching the sacrificial layer 142 and depositing the conductor layer 143. The spacing between adjacent layers 141 is relatively narrow and the distance between the GLSs 170 is relatively long. If GLS 161 were not present, then the etching of layer 142 would have to penetrate the narrow and long spaces between layers 141. Thereafter, some portions of layer 142 horizontally in the middle region may not be completely etched away. If portions of layer 142 are not etched away, portions of conductor layer 143 may not be deposited or may not be properly deposited. Since the conductor layer 143 functions as a gate electrode of the NAND memory cell, an incomplete gate electrode may cause malfunction of the NAND memory cell. Since the GLS 161 is configured to be in the vicinity of the middle region between the GLS 170, the length of the narrow and long space between the layers 141 is divided into two. As such, layer 142 can be completely etched away and conductor layer 143 or gate electrode can be properly deposited.

After forming the conductor layer 143, a first array of common source electrodes (ACS)190 and a second ACS 191 may be fabricated. First, an electrically insulating layer may be deposited on the sidewalls and bottom surfaces of GLS 161 and GLS 170 by CVD, PVD, ALD, or a combination of two or more of these processes. Thereafter, a dry etch process or a combination of dry and wet etch processes may be performed that will remove portions of layer 192 at the bottom of GLS 161 and GLS 170, thereby exposing portions of layer 184.

Thereafter, other processes are performed to form the first ACS 190 and the second ACS 191 in the GLS 161 and the GLS 170. For example, a conductive layer 193 comprising a conductive material such as titanium nitride, W, Co, Cu, Al, doped silicon, or silicide may be deposited on the silicon oxide layer 192, and the conductive layer 193 electrically contacts the layer 184 at the bottom of the GLS 161 and the GLS 170. Thereafter, GLS 161 and GLS 170 may be filled with a conductive material 194 (e.g., doped polysilicon). That is, in some embodiments, the first ACS 190 and the second ACS 191 may each include an isolation layer deposited on sidewalls of the GLS, a conductive material deposited on the isolation layer, and a conductive material filling the GLS. The isolation layer insulates the first ACS 190 and the second ACS 191 from the conductor layer 143. In some embodiments, the first ACS 190 and the second ACS 191 may each include an isolation layer deposited on the sidewalls of the GLS and one or more conductive materials deposited on the isolation layer and filling the GLS. As shown in fig. 12, after the first ACS 190 and the second ACS 191 are formed, they become electrically conductive channels that extend through the layer stack 140 and electrically contact the layer 184.

Since the first ACS 190 is formed by filling the GLS 161 in the first GLS area 160, the first ACS 190 may also have a cylindrical or columnar shape. Similarly, the second ACS 191 may have the same shape as the GLS 170 or a similar shape. As such, the first ACS 190 is arranged in the same manner as the GLS 161 and the second ACS 191 is arranged in the same manner as the GLS 170. That is, the first ACS 190 extends and is spaced apart in a non-continuous manner along the X-direction, and the second ACS 191 extends from the left side to the right side of the memory block area 101 in a continuous manner along the X-direction. The first ACS 190 are spaced apart in a direction parallel to the second ACS 191 and are in an intermediate portion of the area 101 between the second ACS 191. In addition, the second ACS 191 divides the NAND memory unit into a plurality of memory block areas (not shown). Each memory block area may be arranged between a pair of the second ACS 191 in the Y-direction and comprises a trench aperture area comprising a non-consecutive first ACS 190. In some other embodiments, the first ACS 190 may extend in a non-continuous manner in both the X-direction and the Y-direction. When the first ACS extends in both the X-direction and the Y-direction, the corresponding fabrication process may remain the same as when the first ACS extends only in the X-direction.

After the first ACS and the second ACS are formed, other fabrication steps or processes are performed to complete the fabrication of the device 100. Details of other fabrication steps or processes are omitted for simplicity.

Fig. 13 shows a schematic flow diagram 200 for fabricating a 3D memory device according to an embodiment of the present disclosure. At 211, a sacrificial layer may be deposited over a top surface of a substrate. The substrate may comprise a semiconductor substrate, for example, a monocrystalline silicon substrate. In some embodiments, a base layer or a capping layer may be deposited on the substrate prior to depositing the sacrificial layer. The base or cover layer may comprise a single layer or a composite layer having multiple layers deposited sequentially over the substrate. In some embodiments, the base layer or capping layer may comprise silicon oxide, silicon nitride, and/or aluminum oxide. In some other embodiments, the sacrificial layer may be deposited without depositing a base layer or a capping layer over the substrate. The sacrificial layer may comprise monocrystalline silicon, polycrystalline silicon, silicon oxide or silicon nitride.

At 212, a layer stack may be deposited over the sacrificial layer. The layer stack includes first stacked layers and second stacked layers alternately stacked. In some embodiments, the first stacked layer may include a first dielectric layer, and the second stacked layer may include a second dielectric layer different from the first dielectric layer. In some embodiments, one of the first dielectric layer and the second dielectric layer is configured as a sacrificial stack layer. In some other embodiments, the first stacked layer and the second stacked layer may include a dielectric layer and a conductive layer, respectively.

At 213, a channel hole may be formed through the layer stack and the sacrificial layer to expose a portion of the substrate. A functional layer and a channel layer may be deposited on sidewalls of each channel hole. Forming the functional layer may include depositing a barrier layer on sidewalls of the trench hole, depositing a charge trapping layer on the barrier layer, and depositing a tunneling insulating layer on the charge trapping layer. The channel layer deposited on the tunneling insulating layer functions as a semiconductor channel and may include a polysilicon layer.

At 214, first and second GLS may be formed that extend vertically through the layer stack and into the sacrificial layer, and that expose portions of the sacrificial layer. The second GLS also extends continuously in the horizontal direction, and divides the NAND memory cells into a plurality of memory block areas. The memory block region includes a channel hole region including a first GLS. The first GLS may be discontinuous along the horizontal direction and may be spaced apart by one or more predetermined distance values. Since the first GLS is discontinuous, the first GLS does not affect the function of the memory block area.

At 215, the sacrificial layer may be etched away and a cavity may be created over the substrate. The cavity exposes a portion of the barrier layer of the functional layer that is within the cavity. If a base layer or a cover layer is deposited on the substrate, the base layer or the cover layer is also exposed in the cavity. Thereafter, the various layers of the functional layer that are sequentially exposed in the cavity, including the barrier layer, the charge trapping layer, and the tunneling insulating layer, are etched away, respectively, by, for example, one or more selective etching processes. As a result, a portion of the functional layer close to the substrate can be removed in the cavity. The base layer or the capping layer, if deposited, may also be etched away during the process of etching the portion of the functional layer or in another selective etching process. Thus, portions of the substrate and side portions of the channel layer are exposed in the cavity.

At 216, selective epitaxial growth may be performed to grow a single crystal silicon layer on the substrate in the cavity and a polysilicon epitaxial layer on the exposed portions (i.e., sidewalls) of the channel layer. During epitaxial growth, the single crystal silicon layer and the polycrystalline silicon layer abut or merge with each other to fill the cavity. Because the first GLS is disposed in the channel hole region, the epitaxial growth rates of single crystal silicon and polycrystalline silicon in the cavity will be more uniform than in the absence of the first GLS. As such, void formation is avoided when filling the cavity.

In some embodiments, the layer stack comprises two dielectric stack layers, and one of the stack layers is sacrificial. The sacrificial stack layers may be etched away at 217, leaving a cavity, which is then filled with a conductive material at 218 to form a conductor layer. The conductive material may include a metal such as W, Co, Cu, or Al. Since the first GLS is disposed in the middle of the channel hole region, the sacrificial stack layer may be completely etched away. Thus, incomplete gate electrode formation can be avoided.

At 219, an oxide layer may be deposited and selectively etched at the first GLS and the second GLS to expose the epitaxial layer filling the cavity. A conductive material such as titanium nitride, W, Cu, Al and/or doped polysilicon may be deposited in the GLS to form a first ACS and a second ACS, respectively, in electrical contact with the epitaxial layer.

Fig. 14 and 15 schematically illustrate a top view and a cross-sectional view of another 3D memory device 300 according to an embodiment of the present disclosure. The cross-sectional view shown in fig. 15 is taken along line CC' of fig. 14. The structure of the 3D memory device 300 may be similar to that of the device 100, but the first GLS region 162 of the device 300 is different from the first GLS region 160 of the device 100.

As shown in fig. 14 and 15, the channel holes 150 of the device 300 are configured to extend in the Z-direction and form an array of certain patterns in the X-Y plane. The NAND memory cells of the 3D memory device 300 may be divided into a plurality of memory block areas (not shown) by the GLS 171. GLS 171 may have the same structure as or a similar structure to that of GLS 170 of device 100. The memory block region 103, which is separated from other memory block regions (not shown), may include a channel hole region 104. The memory block area 103 may be disposed between a pair of GLSs 171. The trench hole region 104 may include a first GLS region 162, and the first GLS region 162 includes a plurality of GLSs 163. GLS 163 extend discontinuously in the X direction and are spaced apart from each other, while GLS 171 extend continuously in the X direction. In contrast to GLS 161 of device 100, which has a circular cross-section in the X-Y plane, GLS 163 has a diamond-shaped cross-section in the X-Y plane. In some embodiments, as in fig. 14, the GLS 163 may have the same shape and size in the trench hole region 104. In some other embodiments, the GLS 163 may have a different shape and/or a different size in the trench hole region 104.

The method of making the 3D NAND memory device 300 may use one or more processes that are the same as or similar to those used for the device 100. For example, one or more deposition processes, one or more etching processes, and/or one or more filling processes used with respect to device 100 may be used in the fabrication of device 300.

For example, as shown in fig. 14 and 15, in fabricating the 3D memory device 300, the layer 120 may be deposited over the top surface of the substrate 110 by CVD, PVD, ALD, or a combination of two or more of these processes. Next, similar to device 100, a sacrificial layer (e.g., polysilicon layer 130) and a layer stack 140 comprising alternating stacked layers 141 and 142 may be sequentially deposited over layer 120. Similar to device 100, stacked layers 141 and 142 of device 300 may also illustratively comprise silicon oxide and silicon nitride, respectively. As shown in fig. 15, similar to device 100, device 300 may also include channel hole 150, functional layer 151, and polysilicon channel layer 155 (semiconductor channel). Functional layer 151 is formed on the sidewalls and bottom surface of channel hole 150 in the same manner as device 100. The functional layer 151 may exemplarily include a silicon oxide layer as a blocking layer 152 deposited on sidewalls and a bottom of the channel hole 150, a silicon nitride layer as a charge trap layer 153 deposited on a surface of the blocking layer 152, and a silicon oxide layer as a tunneling insulation layer 154 deposited on a surface of the charge trap layer 153. A polysilicon channel layer 155 may be deposited on a surface of the tunnel insulation layer 154. The channel hole 150 may be filled with a dielectric material 156.

Next, GLS 163 and GLS 171 may be formed. As shown in FIG. 14, the cross-section of GLS 163 has a diamond shape in the X-Y plane. Similar to device 100, GLS spacers may be deposited and selectively etched to expose sacrificial layer 130. Next, portions of the sacrificial layer 130, the layer 120, and the functional layer 151 near the substrate may be etched, which exposes side portions of the channel layer 155 and the substrate in the cavity. After exposing the side portions of the channel hole 155 and the substrate, selective epitaxial growth may be performed to grow an epitaxial layer filling the cavity. These epitaxial layers electrically contact exposed side portions of the channel layer 155. Since the GLS 163 is configured as an intermediate region between the GLS 171, formation of voids can be prevented when growing an epitaxial layer filling the cavity.

Thereafter, similar to device 100, stack layer 142 may be etched away and stacked layer 142 may be replaced with a conductor layer (e.g., a W layer). The conductor layer is configured as a word line of the 3D memory device 300, and the channel layer 155 is configured as a bit line. Thereafter, the epitaxial layer may be exposed at the bottom of GLS 163 and GLS 171 using a deposition and etching process of an oxide layer. Electrically conductive material may be deposited in GLS 163 and GLS 171 to form the first ACS and the second ACS. The first ACS and the second ACS each extend through the layer stack 140 and electrically contact the epitaxial layers. After which other fabrication steps or processes are performed to complete the fabrication of the device 300.

Since the GLS 163 has a diamond-shaped cross-section in the X-Y plane, the first ACS formed in the GLS 163 also has a diamond-shaped cross-section in the X-Y plane. Further, similar to the device 100, the first ACS extends in a discontinuous manner and is spaced apart from each other along the X-direction, and the second ACS extends in a continuous manner along the X-direction. Further, the second ACS divides the NAND memory unit into a plurality of memory block areas. Each memory block area may be arranged between a pair of the second ACS in the Y direction. The memory block may comprise a channel hole region comprising the non-contiguous first ACS. In some other embodiments, the GLS 163, and thus the first ACS, may extend in a non-continuous manner in both the X-direction and the Y-direction, respectively.

Fig. 16 schematically illustrates a top view of another 3D memory device 400 according to an embodiment of the present disclosure. The cross-sectional view of device 400 is omitted for simplicity. The structure of the 3D memory device 400 may be similar to that of the device 100 and the device 300, but the first GLS region 164 of the device 400 is different from the first GLS region 160 of the device 100 and the first GLS region 162 of the device 300.

Similar to devices 100 and 300, channel holes 150 of device 400 are configured to extend in the Z-direction and are patterned in the X-Y plane. As shown in fig. 16, the NAND memory cell of the 3D memory device 400 may be divided into a plurality of memory block areas (not shown) by the GLS 172. The GLS 172 may have the same structure as or a similar structure to the GLS 170 of the device 100 or the GLS 171 of the device 300. A memory block area 105, which is separated from other memory block areas (not shown), may be disposed between the pair of GLSs 172, and the memory block area 105 includes the trench hole area 106. The tunnel region 106 may include a first GLS region 164, and the first GLS region 164 may include a plurality of GLSs 165 located near an intermediate region between the GLSs 172. In some embodiments, the GLS 165 may have the same shape and size as the GLS 161 of the device 100. In some other embodiments, the GLS 165 may have a different shape or different size than the shape and size of the GLS 161, such as a square shape, an oval shape, or other shapes.

Further, in contrast to devices 100 and 300 where GLS 161 or GLS 163 form a single row extending in the X-direction, GLS 165 may form two rows each extending in the X-direction. In a row, the GLSs 165 extend discontinuously and are spaced apart from each other. In some embodiments, adjacent GLSs 165 may be spaced apart by a constant distance. In some other embodiments, adjacent GLSs 165 may be spaced apart by different values of distance. In some embodiments, two rows of GLSs 165 may be arranged near the middle region between the GLSs 172 and spaced apart by a predetermined distance.

In some other embodiments, more than two rows of GLSs 165 may be arranged near the middle region between the GLSs 172. As such, the GLS 165 may be configured to extend in both the X-direction and the Y-direction. That is, the GLS 165 may be configured to form an appropriate pattern in the X-Y plane in the memory block area 105 or between the GLS 172. In some embodiments, as in fig. 16, the GLS 165 may extend a longer range in the X-direction than in the Y-direction.

Because adjacent GLSs 165 are spaced apart at least some distance, the gate electrode of each NAND memory cell in the memory block area 105 is electrically connected. As such, the first GLS region 164 or GLS 165 may be considered to divide the trench hole region 106 into three electrically connected portions. Thus, the function of the memory block area 105 is not affected.

When more than one row of GLSs 165 are formed between a pair of GLSs 172, the process for etching the sacrificial layers (e.g., layers 130 and 142 of device 100) may become faster and more complete. Similarly, the growth rate of the epitaxial layer in the cavity (e.g., the growth rate of layer 184 in cavity 181 of device 100) may become more uniform, thereby preventing the formation of voids. Further, in some other embodiments, the distance between the GLS 172 may increase in the Y direction when more channel holes are arranged in the memory block region. The distance between the GLSs 172 may become so large that a row of GLSs 165 may not be sufficient to prevent the formation of voids in the selective epitaxial growth in the cavity (e.g., the growth of layer 184 in cavity 181 of device 100) or to prevent incomplete etching of the sacrificial stack layers (e.g., layer 142 of device 100). Thus, in some embodiments, multiple rows of GLS 165 may be desirable to avoid void formation and incomplete etching.

Similar to the device 100 and the device 300, but not shown in the figures, the device 400 may comprise a substrate 100 and a layer stack 140 disposed above the substrate. The device 400 may also include a functional layer 151, a channel layer 155, a first ACS, and a second ACS extending vertically through the layer stack 140 in the Z-direction. The description of such a structure will be omitted or not repeated in detail.

The method of fabricating the 3D NAND memory device 400 may use one or more processes that are the same as or similar to those used for the devices 100 and 300. For example, one or more deposition processes, one or more etching processes, and/or one or more filling processes used with respect to devices 100 and 300 may be used in the fabrication of device 400. The description of such a fabrication process will be omitted or not repeated in detail.

Fig. 17, 18, 19, and 20 schematically illustrate top and cross-sectional views of another 3D memory device 500 according to an embodiment of the present disclosure. The cross-sectional view shown in fig. 18 is taken along line DD' of fig. 17. The cross-sectional view shown in fig. 19 is taken along line EE' of fig. 20. The structure of the 3D memory device 500 may be similar to that of the device 100, the device 300, and/or the device 400, but the first GLS region 166 of the device 500 is different from the first GLS region 160 of the device 100, the first GLS region 162 of the device 300, and the first GLS region 164 of the device 400.

Similar to devices 100, 300, and 400, channel holes 150 of device 500 are configured to extend in the Z-direction and form a pattern in the X-Y plane. As shown in fig. 17, the NAND memory cell of the 3D memory device 500 may be divided into a plurality of memory block areas (not shown) by the GLS 173. GLS 173 may have the same structure as or a similar structure to that of GLS 170 of device 100 or GLS 171 of device 300. The memory block region 107 may include a channel hole region 108. The memory block area 107 may be disposed between a pair of GLSs 173 and separated from other memory block areas (not shown) by the GLSs 173. The trench hole region 108 may include a first GLS region 166, and the first GLS region 166 may include a plurality of GLSs 167 located near a middle region between the GLSs 173. Similar to devices 100, 300, and 400, GLS 167 extends continuously in the Z-direction and discontinuously in the X-direction. The cross-section of the GLS 167 may have various shapes in the X-Y plane, for example, a rectangular shape as shown in fig. 17.

Similar to device 400, but unlike devices 100 and 300, GLS 167 may form two rows that run parallel to GLS 173 and along the X-direction. In some embodiments, two rows of GLSs 167 may be arranged near the middle region between the GLSs 173 and spaced apart by a predetermined distance. In each row, the GLS 167 extend discontinuously and are spaced apart from each other. In some embodiments, adjacent GLSs 167 may be spaced apart by a constant distance along the X-direction. In some embodiments, adjacent GLSs 167 may be separated by different values of distance along the X-direction. In some other embodiments, more than two rows of GLSs 167 may be arranged near the middle region between the GLSs 173. As such, the GLS 167 may be configured to extend in both the X and Y directions, and form a 2D pattern in the X-Y plane in the memory block region 107 or between the GLS 173. In some embodiments, the GLS 167 may extend a longer range in the X-direction than in the Y-direction.

Since the adjacent GLSs 165 are spaced apart by at least a certain distance in the X direction, the Y direction, or another direction between the X direction and the Y direction, the gate electrodes of each NAND memory cell in the memory block area 107 are electrically connected. As such, the first GLS region 166 or GLS 167 may be considered as a portion that divides the trench hole region 108 into three electrical connections. Thus, the function of the memory block area 107 is not affected by the GLS 167.

Similar to device 400, when more than one row of GLSs 167 is disposed between a pair of GLSs 173, the process for etching the sacrificial layer (e.g., layer 130 or layer 142 of device 100) may become faster and more complete than it would be if GLS 167 were not present. Furthermore, the growth rate of the epitaxial layer in the cavity (e.g., the growth rate of layer 184 in cavity 181 of device 100) may become more uniform, thereby preventing the formation of voids. Furthermore, in some other embodiments, when more channel holes are disposed between GLSs 173, the distance between GLSs 173 may be increased accordingly. The distance between GLS 173 may become so large that a row of GLS 167 may not be sufficient to prevent the formation of voids in the selective epitaxial growth in the cavity (e.g., layer 184 in cavity 181 of device 100) or to prevent incomplete etching of the sacrificial stack layers (e.g., layer 142 of device 100). Thus, in some embodiments, multiple rows of GLS 167 may be needed to avoid void formation and incomplete etching.

Similar to device 100, device 300, and device 400, as shown in fig. 19, device 500 may include substrate 110, epitaxial layer 185, layer stack 140, first ACS195, and second ACS 196.

The method of fabricating the 3D NAND memory device 500 may use one or more processes that are the same as or similar to those used for the devices 100, 300 and 400. For example, one or more deposition processes, one or more etching processes, and/or one or more filling processes used with respect to devices 100, 300, and 400 may be used in the fabrication of device 500.

As shown in fig. 18 and 19, in fabricating the 3D memory device 500, the capping layer 120 may be deposited over the top surface of the substrate 110 by CVD, PVD, ALD, or a combination of two or more of these processes. Next, similar to devices 100, 300, and 400, a sacrificial layer (e.g., polysilicon layer 130) and a layer stack 140 comprising alternating stack layers 141 and 142 may be sequentially deposited over capping layer 120. Similar to devices 100, 300, and 400, stacked layers 141 and 142 of device 500 may also illustratively comprise silicon oxide and silicon nitride, respectively. As shown in fig. 17 and 18, the device 500 may further include a channel hole 150, a functional layer 151, and a polysilicon channel layer 155 (semiconductor channel), similar to the devices 100, 300, and 400. Functional layer 151 is formed on the sidewalls and bottom surface of channel hole 150 in the same manner as device 100. The functional layer 151 may exemplarily include a silicon oxide layer as a blocking layer 152 deposited on sidewalls and a bottom of the channel hole 150, a silicon nitride layer as a charge trap layer 153 deposited on a surface of the blocking layer 152, and a silicon oxide layer as a tunneling insulation layer 154 deposited on a surface of the charge trap layer 153. A polysilicon channel layer 155 may be deposited on a surface of the tunnel insulation layer 154. The channel hole 150 may be filled with a dielectric material 156.

Next, GLS 167 and GLS 173 may be formed. As shown in fig. 17, the cross-section of the GLS 167 has a rectangular shape in the X-Y plane. Thereafter, similar to device 100, GLS spacers 168 may be deposited and selectively etched to expose sacrificial layer 130, as shown in fig. 18. Next, the sacrificial layer 130, the capping layer 120, portions of the GLS spacer 168, and portions of the functional layer 151 adjacent to the substrate may be etched away, which exposes side portions of the channel layer 155 and the substrate in the cavity. After exposing the side portions of the channel layer 155 and the substrate, selective epitaxial growth may be performed to grow an epitaxial single crystal silicon layer and a polysilicon layer on the substrate and the side portions of the channel layer 155. Epitaxial growth fills the cavity and an epitaxial layer 185 is formed. The epitaxial layer 185 is in electrical contact with the side portion of the channel layer 155. Since two rows of GLS 167 are disposed in the middle region between the GLS 173, the formation of voids can be prevented when growing the epitaxial layer 185 filling the cavity.

Thereafter, similar to the devices 100, 300, and 400, the remaining portions of the GLS spacers 168 and the sacrificial stack layer 142 may be etched away, and the layer 142 may be replaced with a conductor layer 143 (e.g., a W layer). Since the two rows of GLSs 167 are configured as the middle region between the GLSs 173, the stacked layer 142 may be completely etched away. Thus, the conductor layer 143 can be correctly deposited without problems caused by incomplete etching of the layer 142. The conductor layer 143 is configured as a word line of the 3D memory device 500, and the channel layer 155 is configured as a bit line. Next, a first ACS195 and a second ACS 196 are fabricated. A deposition process may be performed to form an electrical isolation layer, e.g., a silicon oxide layer 197, on the sidewalls and bottom surface of the GLS 167 and the GLS 173. Layer 197 may be selectively etched to expose epitaxial layer 185 at the bottom of GLS 167 and GLS 173.

Next, other processes may be performed to form the first ACS195 and the second ACS 196 in the GLS 167 and the GLS 173. As shown in fig. 19, a conductive layer 198 comprising a conductive material such as titanium nitride, W, Co, Cu, Al, doped silicon or suicide may be deposited to cover the silicon oxide layer 197 and the conductive layer 198 electrically contacts the layer 185 at the bottom of the GLS 167 and GLS 173. These GLSs may then be filled with a conductive material 199 (e.g., doped polysilicon). Fig. 20 schematically illustrates a top view after filling the GLS to form the first ACS195 and the second ACS 196. As shown in fig. 19, after the first ACS195 and the second ACS 196 are formed, they become conductive channels that extend through the layer stack 140 and electrically contact the epitaxial layer 185. After which other fabrication steps or processes are performed to complete the fabrication of the device 500.

Since the GLS 167 has a rectangular cross-section in the X-Y plane, the first ACS195 formed in the GLS 167 also has a rectangular cross-section in the X-Y plane. Further, similar to the devices 100, 300, and 400, the first ACS195 extends and is spaced apart in a non-continuous manner along the X-direction, and the second ACS 196 extends from the left side to the right side of the memory block area 107 in a continuous manner along the X-direction. In addition, the second ACS 196 may divide the NAND memory cells of the device 500 into a plurality of memory block areas (not shown). Each memory block area (e.g., memory block area 107) may be disposed between a pair of second ACS 196 in the Y-direction. Further, each memory block region may include a channel hole region including the non-continuous first ACS 195. As shown in fig. 17 and 19, the first ACS 196 may extend in a non-continuous manner in both the X-direction and the Y-direction, respectively. That is, in the memory block area 107 or between the pair of second ACS 196, the first ACS195 may be configured to form a pattern, e.g., a 2D pattern, in the X-Y plane. Since the second ACS 196 extends continuously in the X-direction from the left side to the right side, the first ACS195 may extend a longer range in the X-direction than in the Y-direction.

In some embodiments, the interval between adjacent first ACS195 in the X direction may be a fixed value. In some other embodiments, the interval between adjacent first ACS195 in the X direction may comprise different values. Further, in some embodiments, the GLS 167 may have the same shape or size in the trench hole region 108. In some other embodiments, each GLS 167 may have a different shape or size in the trench hole region 108. Accordingly, in some embodiments, the first ACS195 may have the same shape or size in the trench hole region 108. In some other embodiments, the first ACS195 may have a different shape or size in the trench hole region 108.

The first ACS195 is arranged to be discontinuous, not in contact with each other, and spaced apart at least a given distance. As such, the spacing between adjacent first ACS195 provides an electrical connection between the NAND memory cells. The first ACS195 may be considered as dividing the trench hole region 108 into three electrically connected portions. Thus, the function of the memory blocks in the memory block area (e.g., area 107) is not affected by the arrangement of the first ACS 195.

Fig. 21 and 22 schematically illustrate top and cross-sectional views of a 3D memory device 500 having additional features according to embodiments of the present disclosure. The cross-sectional view shown in fig. 22 is taken along line FF' of fig. 21. The additional feature is a Top Select Gate (TSG) cut. Area 1951 represents a TSG cut and is shown in dashed lines in fig. 21 and 22. As shown in fig. 21 and 22, the TSG cut has a width narrower than that of the first ACS195 in the Y-direction and continuously extends from the left side to the right side of the trench hole region 108 in the X-direction. In the vertical direction (i.e., Z-direction), the TSG cut extends within a limited range and only partially through the layer stack 140. Thus, the area 1951 and the first ACS195 partially overlap in the trench hole area 108 and the memory block area 107.

In some embodiments, a row of discontinuous first ACS195 may be configured between the second ACS 196 along with the TSG cuts. For example, if there are thirteen rows of channel holes 150 along the Y-direction between the second ACS 196, in some embodiments, the seventh row of channel holes 150 from the second ACS 196 may be used to form a row of the first ACS 195. Thus, six rows of channel holes 150 are arranged on each side of the first ACS195 of the row. Further, multiple rows of the first ACS195 may be disposed between the second ACS 196 along the Y-direction along with multiple TSG cuts. For example, in some embodiments, when there are 7N 1 rows of channel holes 150 (where N is an integer greater than 2), the seventh row from ACS 196 may be used to form the first ACS195 of the first row, and the fourteenth row may be used to form the first ACS195 of the second row, and so on. In such a case, there are six rows of channel holes 150 between the first ACS195 of two adjacent rows or between the ACS 196 and the ACS195 of an adjacent row.

As shown in fig. 21 and 22, the smaller spacing between adjacent ACS195 in the X direction means a larger area of ACS195 in the X-Z plane and is therefore desirable for the etching and filling processes described above. However, if the spacing between adjacent ACS195 becomes too small, reliability problems may occur because adjacent ACS195 may contact each other. Thus, the length of the ACS195 in the X direction should be optimized.

By using the disclosed memory structure and method, discontinuous GLSs are formed between consecutive GLSs in the memory block area without affecting the functionality of the memory block. The discontinuous GLS improves the selective epitaxial growth in the cavity. Thus, formation of voids can be prevented, and a leakage problem of current can be avoided. The non-continuous GLS also enhances the etching of the sacrificial stack layer, which may prevent incomplete gate electrodes. As such, the quality and reliability of the 3D memory device may be improved.

While the principles and implementations of the present disclosure have been described in this specification using specific embodiments, the foregoing description of the embodiments is merely intended to facilitate an understanding of the present disclosure. Furthermore, features from different embodiments described above can be combined to form additional embodiments. Modifications to the described embodiments and the scope of the application will occur to those skilled in the art in light of the teachings of this disclosure. Accordingly, the contents of the specification should not be construed as limiting the present disclosure.