Non-volatile memory and method of fabricating the same

The integration of an oxide semiconductor layer as the channel layer in non-volatile memory structures addresses the challenges of speed and leakage current, enhancing performance through increased electron mobility and reduced leakage.

TWI931995BActive Publication Date: 2026-07-11POWERCHIP SEMICON MFG CORP
0 Cites 0 Cited by

Patent Information

Application Number
TW114101267
Authority / Receiving Office
TW · TW
Patent Type
Patents
Current Assignee / Owner
Filing Date
2025-01-13
Publication Date
2026-07-11
Estimated Expiration
2045-01-12

AI Technical Summary

Technical Problem

Existing non-volatile memory technologies face challenges in achieving high read and write speeds while minimizing leakage current.

Method used

The use of an oxide semiconductor layer, such as InMO3(ZnO)m, as the channel layer in a non-volatile memory structure, which includes a gate stack and spacer walls, enhances electron mobility and reduces leakage current.

Benefits of technology

This configuration improves read and write speeds and reduces leakage current, thereby increasing device efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure IMG-2_DRAW_114101267-A0305-14-0001-1
    Figure IMG-2_DRAW_114101267-A0305-14-0001-1
  • Figure IMG-2_DRAW_114101267-A0305-14-0001-2
    Figure IMG-2_DRAW_114101267-A0305-14-0001-2
  • Figure IMG-2_DRAW_114101267-A0305-14-0002-3
    Figure IMG-2_DRAW_114101267-A0305-14-0002-3
Patent Text Reader

Abstract

This invention provides a non-volatile memory and a method for forming the same, wherein the non-volatile memory includes: a substrate; an oxide semiconductor layer on the substrate; a tunneling dielectric layer on the oxide semiconductor layer; a gate stack on the tunneling dielectric layer; an erase gate on the tunneling dielectric layer; a second inter-gate dielectric layer between the gate stack and the erase gate; and a source region and a drain region on the substrate and located on opposite sides of the oxide semiconductor layer. The gate stack includes: a floating gate on the tunneling dielectric layer; a first inter-gate dielectric layer on the floating gate; and a control gate on the first inter-gate dielectric layer.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to a non-volatile memory, and more particularly to a non-volatile memory and a method of manufacturing the same. Prior Technology

[0002] With the widespread adoption of consumer electronics such as smartphones, memory cards, and USB flash drives, and the maturity of manufacturing processes, the application of non-volatile memory, such as flash RAM, read-only memory (ROM), erasable programmable read-only memory (EPROM), and electrically erasable programmable read-only memory (EEPROM), has grown significantly.

[0003] However, as the demand for various electronic products increases, there are also higher requirements for the read and write speeds of non-volatile memory, and it is also expected that the leakage current of non-volatile memory can be reduced. Summary of the Invention

[0004] This invention provides a non-volatile memory and a method for forming it, so as to improve the read and write speed of the non-volatile memory and reduce its leakage current.

[0005] This invention proposes a non-volatile memory, comprising: a substrate; an oxide semiconductor layer on the substrate; a tunneling dielectric layer on the oxide semiconductor layer; and a gate stack on the tunneling dielectric layer, wherein the gate stack includes: a floating gate on the tunneling dielectric layer, a first inter-gate dielectric layer on the floating gate, and a control gate on the first inter-gate dielectric layer; an erase gate on the tunneling dielectric layer; a second inter-gate dielectric layer between the gate stack and the erase gate; and a source region and a drain region on the substrate and located on opposite sides of the oxide semiconductor layer.

[0006] According to one embodiment of the present invention, the oxide semiconductor layer includes In-Sn-Ga-Zn-O, In-Ga-Zn-O, In-Sn-Zn-O, In-Al-Zn-O, Sn-Ga-Zn-O, Al-Ga-Zn-O, Sn-Al-Zn-O, In-Zn-O, Sn-Zn-O, Al-Zn-O, Zn-Mg-O, Sn-Mg-O, In-Mg-O, In-O, Sn-O, or Zn-O.

[0007] According to one embodiment of the present invention, the oxide semiconductor layer comprises a material represented by the chemical formula InMO 3(ZnO)m, wherein M is one or more selected from Ga, Al, Mn and Co, and m>0.

[0008] According to one embodiment of the present invention, the thickness of the oxide semiconductor layer is 2 nm to 200 nm.

[0009] According to one embodiment of the present invention, the oxide semiconductor layer is a channel layer of a non-volatile memory.

[0010] According to one embodiment of the present invention, it further includes: a first spacer wall located on the sidewall of the gate stack; and a second spacer wall located on the sidewall of the erased gate.

[0011] According to one embodiment of the present invention, the top surface of the control gate and the top surface of the erase gate are coplanar.

[0012] This invention proposes a method for manufacturing a non-volatile memory, comprising: providing a semiconductor substrate; providing a substrate; forming an oxide semiconductor layer on the substrate; forming a tunneling dielectric layer on the oxide semiconductor layer; forming a gate stack on the tunneling dielectric layer, wherein the gate stack includes: a floating gate located on the tunneling dielectric layer, a first inter-gate dielectric layer located on the floating gate, and a control gate located on the first inter-gate dielectric layer; forming an erase gate located on the tunneling dielectric layer; forming a second inter-gate dielectric layer between the gate stack and the erase gate; and forming a source region and a drain region located on the substrate and on opposite sides of the oxide semiconductor layer.

[0013] According to one embodiment of the present invention, the step of forming an oxide semiconductor layer on a substrate includes: depositing an oxide semiconductor material on the substrate; and patterning the oxide semiconductor material to define the oxide semiconductor layer.

[0014] According to one embodiment of the present invention, the step of depositing oxide semiconductor material on a substrate includes physical vapor deposition (PVD), chemical vapor deposition (CVD), or atomic layer deposition (ALD).

[0015] According to one embodiment of the present invention, the oxide semiconductor layer includes In-Sn-Ga-Zn-O, In-Ga-Zn-O, In-Sn-Zn-O, In-Al-Zn-O, Sn-Ga-Zn-O, Al-Ga-Zn-O, Sn-Al-Zn-O, In-Zn-O, Sn-Zn-O, Al-Zn-O, Zn-Mg-O, Sn-Mg-O, In-Mg-O, In-O, Sn-O, or Zn-O.

[0016] According to one embodiment of the present invention, the oxide semiconductor layer comprises a material represented by the chemical formula InMO3(ZnO)m, wherein M is one or more selected from Ga, Al, Mn and Co, and m>0.

[0017] According to one embodiment of the present invention, the thickness of the oxide semiconductor layer is 2 nm to 200 nm.

[0018] According to one embodiment of the present invention, the oxide semiconductor layer is a channel layer of a non-volatile memory.

[0019] According to one embodiment of the present invention, it further includes: forming a first spacer wall on the sidewall of the gate stack; and forming a second spacer wall on the sidewall of the wiped-out gate.

[0020] According to one embodiment of the present invention, the top surface of the control gate and the top surface of the erase gate are coplanar.

[0021] Based on the above, in the non-volatile memory and its manufacturing method of the present invention, an oxide semiconductor layer is used as the channel layer. Based on the characteristics of the oxide semiconductor layer material itself, the electrons in the non-volatile memory can have a high mobility, thus increasing the read and write speed of the non-volatile memory. Furthermore, using an oxide semiconductor layer as the channel layer of the non-volatile memory can also reduce the non-volatile memory and improve the device efficiency.

[0022] To make the above features and advantages of the present invention more apparent and understandable, specific embodiments are described below in conjunction with the accompanying drawings for detailed explanation. Simple Explanation of the Diagram

[0023] Figures 1 to 10 are flow cross-sectional diagrams of a method for manufacturing a non-volatile memory according to an embodiment of the present invention. Implementation

[0024] The following examples are described in detail with reference to the accompanying drawings, but the examples provided are not intended to limit the scope of the invention.

[0025] For ease of understanding, the same components will be indicated by the same symbols in the following descriptions, and will not be repeated in the following paragraphs.

[0026] Furthermore, the accompanying drawings are for illustrative purposes only and are not drawn to their original dimensions. In fact, for clarity of explanation, the dimensions of various features may be increased or decreased as appropriate.

[0027] Furthermore, for ease of description, the terms "above," "below," and similar spatial relative terms are used herein to describe the relative relationship between one component and another as shown in the diagrams, and are not intended to limit the invention. Therefore, it should be understood that "above" and "below" can be used interchangeably, and when a component is located "above" another component, the component can be placed directly on the other component, or there may be an intermediate component. On the other hand, when a component is said to be placed "directly" on another component, there is no intermediate component between them.

[0028] The terminology used herein is for illustrative purposes only and is not intended to limit the scope of this disclosure. In this context, the singular form includes the plural form unless the context otherwise requires.

[0029] Please refer to Figure 1. First, a substrate 100 is provided, and then an oxide semiconductor layer 110 is formed on the substrate 100.

[0030] In some embodiments, the substrate 100 may include materials such as silicon and glass, but is not limited thereto; any substrate 100 used to manufacture the non-volatile memory 10 may be used.

[0031] In some embodiments, the oxide semiconductor layer 110 may include metal oxides, such as quaternary metal oxides In-Sn-Ga-Zn-O; ternary metal oxides In-Ga-Zn-O, In-Sn-Zn-O, In-Al-Zn-O, Sn-Ga-Zn-O, Al-Ga-Zn-O, Sn-Al-Zn-O; binary metal oxides In-Zn-O, Sn-Zn-O, Al-Zn-O, Zn-Mg-O, Sn-Mg-O, In-Mg-O; and monometallic metal oxides In-O, Sn-O, Zn-O, etc.

[0032] In addition, the oxide semiconductor layer 110 may include a material represented by the chemical formula InMO 3(ZnO) m, wherein M is one or more metal elements selected from Ga, Al, Mn and Co. For example, M may be Ga alone, or a combination of two metal elements such as Ga and Al, Ga and Mn, or Ga and Co, and m>0.

[0033] Furthermore, in materials represented by InMO 3(ZnO) m (m>0), materials where M is Ga are often referred to as IGZO (Indium Gallium Zinc Oxide).

[0034] In some embodiments, various suitable deposition methods can be used to form oxide semiconductor materials on the substrate 100, such as physical vapor deposition (PVD), chemical vapor deposition (CVD), or atomic layer deposition (ALD), but are not limited thereto; any method that can be used to form oxide semiconductor materials can be used in this invention. Then, patterning is performed using methods such as photolithography to define the desired cell area and shape on the oxide semiconductor layer 110.

[0035] The cell size and shape defined in the oxide semiconductor layer 110 must be able to accommodate the main components of the non-volatile memory 10 such as the floating gate FG, control gate CG, and erase gate EG that are subsequently formed, but the requirements for component density must also be taken into account, and it cannot be expanded indefinitely.

[0036] Furthermore, the thickness of the oxide semiconductor layer 110 varies depending on factors such as the material used for the oxide semiconductor layer 110 and the size of the non-volatile memory 10 to be formed. Generally, the thickness of the oxide semiconductor layer 110 can be approximately 2 nm to 200 nm, preferably 2 nm to 5 nm. However, this thickness range is a result of matching the current size of the non-volatile memory 10. As the size of the non-volatile memory 10 changes in the future, the thickness of the oxide semiconductor layer 110 must also be changed accordingly.

[0037] The oxide semiconductor layer 110 can be used as a channel layer for the non-volatile memory 10 as shown in FIG10. Based on the characteristics of the oxide semiconductor layer 110 material itself, the electrons in the non-volatile memory 10 can have a high mobility, thus increasing the read and write speed of the non-volatile memory 10. Furthermore, using the oxide semiconductor layer 110 as a channel layer for the non-volatile memory 10 can also reduce leakage current and improve device efficiency.

[0038] Next, referring to Figure 2, a tunneling dielectric layer 122 is formed on the oxide semiconductor layer 110.

[0039] In some embodiments, the tunneling dielectric layer 122 may include an oxide layer, but is not limited thereto.

[0040] Next, please refer to Figures 2 and 3 simultaneously. A gate stack G is formed on the tunneling dielectric layer 122, wherein the gate stack G includes a floating gate FG located on the tunneling dielectric layer 122, a first inter-gate dielectric layer 126 located on the floating gate FG, and a control gate CG located on the first inter-gate dielectric layer 126.

[0041] In some embodiments, the step of forming a gate stack G on the tunneling dielectric layer 122 may include: sequentially forming an electron access layer 124, a dielectric layer 126, and a first polysilicon layer 128 on the tunneling dielectric layer 122, as shown in FIG2; next, using methods such as photolithography, patterning the first polysilicon layer 128, the dielectric layer 126, and the electron access layer 124, so as to sequentially form a control gate CG, a first inter-gate dielectric layer 126, and a floating gate FG on the tunneling dielectric layer 122, thereby completing the definition of the gate stack G, as shown in FIG3.

[0042] In some embodiments, the electronic access layer 124 may include any electronic access material suitable for non-volatile memory 10, such as silicon nitride (Si3N4) or polycrystalline silicon, but is not limited thereto, to be used as the floating gate FG of non-volatile memory 10.

[0043] The dielectric layer 126 may include any material suitable for the floating gate FG and control gate CG of the non-volatile memory 10, so as to give the non-volatile memory 10 a smaller leakage current and better data retention capability, thereby improving the reliability of the non-volatile memory 10. In some embodiments, the dielectric layer 126 may include an oxide layer, a nitride layer, or a material combining multiple oxide and nitride layers, such as an oxide-nitride-oxide (ONO) layer or an oxide-nitride-oxide-nitride (ONON) layer.

[0044] After the cell definition of the gate stack G, the first polycrystalline silicon layer 128 forms the control gate CG of the non-volatile memory 10.

[0045] Next, as shown in Figure 4, a first oxide layer 130 can be formed on the top and side surfaces of the gate stack G by deposition or tempering.

[0046] Then, a second polycrystalline silicon layer 140 is deposited on the tunneling dielectric layer 122 and the gate stack G.

[0047] In some embodiments, the second polysilicon layer 140 can completely cover the gate stack G, that is, the thickness of the deposited second polysilicon layer 140 can exceed the thickness of the gate stack G. Then, the excess second polysilicon layer 140 in the vertical direction is removed by a planarization method such as chemical mechanical polishing (CMP). Since there is a certain degree of etch selectivity between the polysilicon and the oxide layer, the removal of the second polysilicon layer 140 can be stopped at a horizontal position with the top surface of the first oxide layer 130. That is, at this time, the top surface of the first oxide layer 130 and the top surface of the second polysilicon layer 140 are coplanar, as shown in FIG5.

[0048] Next, the second polysilicon layer 140 is patterned by means of photolithography to define the size of the erase gate EG of the non-volatile memory 10 on the tunneling dielectric layer 122, as shown in Figure 6.

[0049] Next, a second oxide layer 150 can be formed on the top and sidewalls of the erased gate EG by means of direct deposition or thermal oxidation.

[0050] If a second oxide layer 150 is formed on the top and sidewalls of the erase gate EG by thermal oxidation, since the erase gate EG is formed by patterning the second polycrystalline silicon layer 140, and its material is polycrystalline silicon, the polycrystalline silicon exposed to oxygen will oxidize into silicon oxide in an oxygen-filled and high-temperature environment. Therefore, in the erase gate EG shown in Figure 6, the outer part of the erase gate EG exposed to oxygen will be thermally oxidized. As a result, the polycrystalline silicon erase gate EG will shrink inward to the erase gate EG shown in Figure 7, and the second oxide layer 150 will be formed on the exposed top and sidewalls of the erase gate EG shown in Figure 6. The degree of shrinkage of the erase gate EG is not as obvious as shown in Figure 7; this exaggerated illustration is for clearer description and illustration.

[0051] Next, as shown in Figure 8, various etching methods can be used to remove the exposed tunneling dielectric layer 122, the first oxide layer 130 on the top surface of the gate stack G, and the second oxide layer 150 on the top surface of the erased gate EG, so as to expose the top surface of the substrate 100, the top surface of the control gate CG, and the top surface of the erased gate EG.

[0052] The first oxide layer 130 remaining after this etching process is formed as a first spacer S1 on the sidewall of the gate stack G and a second intergate dielectric layer 132 between the gate stack G and the erased gate EG; and the second oxide layer 150 remaining after this etching process is formed as a second spacer S2 on the sidewall of the erased gate EG.

[0053] In some embodiments, dry etching can be used to remove the exposed tunneling dielectric layer 122, the first oxide layer 130 on the top surface of the gate stack G, and the second oxide layer 150 on the top surface of the erase gate EG, so as to expose the top surface of the substrate 100, the top surface of the control gate CG, and the top surface of the erase gate EG. The first oxide layer 130 remaining after this etching process is formed as a first spacer S1 on the sidewall of the gate stack G and a second intergate dielectric layer 132 between the gate stack G and the erase gate EG. The second oxide layer 150 remaining after this etching process is formed as a second spacer S2 on the sidewall of the erase gate EG.

[0054] In some embodiments, wet etching can be used to remove the exposed tunneling dielectric layer 122, the first oxide layer 130 on the top surface of the gate stack G, and the second oxide layer 150 on the top surface of the erased gate EG, so as to expose the top surface of the substrate 100, the top surface of the control gate CG, and the top surface of the erased gate EG. The first oxide layer 130 remaining after this etching process is formed as a first spacer S1 on the sidewall of the gate stack G and a second intergate dielectric layer 132 between the gate stack G and the erased gate EG; and the second oxide layer 150 remaining after this etching process is formed as a second spacer S2 on the sidewall of the erased gate EG. Especially when the tunneling dielectric layer 122 is an oxide, since the tunneling dielectric layer 122, the first oxide layer 130 and the second oxide layer 150 to be removed are all oxides, they have a relatively high etch selectivity relative to the oxide semiconductor layer 110. Therefore, the damage to the oxide semiconductor layer 110, which serves as the channel layer, can be avoided in this wet etching process.

[0055] In some embodiments, the top surface of the control gate CG and the top surface of the erase gate EG can be coplanar, as shown in Figure 8.

[0056] Next, as shown in FIG9, a conductive layer 160 may be formed on the substrate 100. This conductive layer 160 may include conductive materials such as titanium nitride (TiN) to reduce the resistance between the contact plug 170 subsequently formed on the substrate 100 and the substrate 100.

[0057] Then, as shown in FIG10, source regions and drain regions S / D are formed on opposite sides of the oxide semiconductor layer 110 on the substrate 100, and a plurality of contact plugs 170 are formed on the top surface of the conductive layer 160, the top surface of the control gate CG and the top surface of the erase gate EG to facilitate the control of the device, and these plurality of contact plugs 170 further include insulating materials such as silicon oxide (not shown).

[0058] In the non-volatile memory 10 of this embodiment, the oxide semiconductor layer 110 can be used as the channel layer of the non-volatile memory 10 as shown in FIG10. Based on the characteristics of the oxide semiconductor layer 110 material itself, the electrons in the non-volatile memory 10 can have a high mobility, thus increasing the read and write speed of the non-volatile memory 10; and, by using the oxide semiconductor layer 110 as the channel layer of the non-volatile memory 10, leakage current can also be reduced and device efficiency can be improved.

[0059] In summary, in the non-volatile memory and its manufacturing method of the present invention, an oxide semiconductor layer is used as the channel layer. Based on the characteristics of the oxide semiconductor layer material itself, the electrons in the non-volatile memory can have a high mobility, thus increasing the read and write speed of the non-volatile memory. Furthermore, using an oxide semiconductor layer as the channel layer of the non-volatile memory can also reduce the non-volatile memory and improve the device efficiency.

[0060] Although the present invention has been disclosed above by way of embodiments, it is not intended to limit the present invention. Anyone skilled in the art can make some modifications and refinements without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

[0061] 10: Non-volatile memory 100: Base 110: Oxide semiconductor layer 122: Tunneling dielectric layer 124: Electronic Access Layer 126: Dielectric layer / First intergate dielectric layer 128: First polycrystalline silicon layer 130: First oxide layer 132: Second intergate dielectric layer 140: Second polycrystalline silicon layer 150: Second oxide layer 160: Conductive layer 170: Contact plug CG: Control Gate EG: Gate removal FG: Floating gate G: Gate Stacking S1: First partition wall S2: Second partition wall S / D: Source Region and Drain Region

Claims

1. A non-volatile memory, comprising: Base; An oxide semiconductor layer is located on the substrate; A tunneling dielectric layer is located on the oxide semiconductor layer; A gate stack is located on the tunneling dielectric layer, wherein the gate stack includes: a floating gate located on the tunneling dielectric layer; a first inter-gate dielectric layer located on the floating gate; and a control gate located on the first inter-gate dielectric layer, wherein the floating gate, the first inter-gate dielectric layer, and the control gate have the same length in a horizontal direction parallel to the surface of the substrate; an erase gate located on the tunneling dielectric layer, wherein the gate stack of the floating gate, the first inter-gate dielectric layer, and the control gate and the erase gate are disposed on the oxide semiconductor layer; a second inter-gate dielectric layer is located between the gate stack and the erase gate; and a source region and a drain region are located on the substrate and on opposite sides of the oxide semiconductor layer.

2. The non-volatile memory as claimed in claim 1, wherein the oxide semiconductor layer comprises In-Sn-Ga-Zn-O, In-Ga-Zn-O, In-Sn-Zn-O, In-Al-Zn-O, Sn-Ga-Zn-O, Al-Ga-Zn-O, Sn-Al-Zn-O, In-Zn-O, Sn-Zn-O, Al-Zn-O, Zn-Mg-O, Sn-Mg-O, In-Mg-O, In-O, Sn-O, or Zn-O.

3. The non-volatile memory as claimed in claim 1, wherein the oxide semiconductor layer comprises a material represented by the chemical formula InMO3(ZnO)m, wherein M is one or more selected from Ga, Al, Mn and Co, and m > 0.

4. The non-volatile memory as claimed in claim 1, wherein the thickness of the oxide semiconductor layer is 2 nm to 200 nm.

5. The non-volatile memory as claimed in claim 1, wherein the oxide semiconductor layer is the channel layer of the non-volatile memory.

6. The non-volatile memory as claimed in claim 1, further comprising: a first spacer wall located on a sidewall of the gate stack; and a second spacer wall located on a sidewall of the erase gate.

7. The non-volatile memory as claimed in claim 1, wherein the top surface of the control gate and the top surface of the erase gate are coplanar.

8. A method for manufacturing a non-volatile memory, comprising: Provide a base; On the substrate where the oxide semiconductor layer is formed; A tunneling dielectric layer is formed on the oxide semiconductor layer; A gate stack is formed on the tunneling dielectric layer, wherein the gate stack includes: a floating gate located on the tunneling dielectric layer; a first inter-gate dielectric layer located on the floating gate; and a control gate located on the first inter-gate dielectric layer, wherein the floating gate, the first inter-gate dielectric layer, and the control gate have the same length in a horizontal direction parallel to the surface of the substrate; an erase gate is formed on the tunneling dielectric layer, wherein the gate stack of the floating gate, the first inter-gate dielectric layer, and the control gate and the erase gate are disposed on the oxide semiconductor layer; a second inter-gate dielectric layer is formed between the gate stack and the erase gate; and a source region and a drain region are formed on the substrate and located on opposite sides of the oxide semiconductor layer.

9. A method for manufacturing a non-volatile memory as claimed in claim 8, wherein the step of forming the oxide semiconductor layer on the substrate comprises: depositing an oxide semiconductor material on the substrate; and patterning the oxide semiconductor material to define the oxide semiconductor layer.

10. A method for manufacturing a non-volatile memory as claimed in claim 8, wherein the step of depositing the oxide semiconductor layer on the substrate includes physical vapor deposition (PVD), chemical vapor deposition (CVD), or atomic layer deposition (ALD).

11. A method for manufacturing a non-volatile memory as claimed in claim 8, wherein the oxide semiconductor layer comprises In-Sn-Ga-Zn-O, In-Ga-Zn-O, In-Sn-Zn-O, In-Al-Zn-O, Sn-Ga-Zn-O, Al-Ga-Zn-O, Sn-Al-Zn-O, In-Zn-O, Sn-Zn-O, Al-Zn-O, Zn-Mg-O, Sn-Mg-O, In-Mg-O, In-O, Sn-O, or Zn-O.

12. A method for manufacturing a non-volatile memory as claimed in claim 8, wherein the oxide semiconductor layer comprises a material represented by the chemical formula InMO3(ZnO)m, wherein M is one or more selected from Ga, Al, Mn and Co, and m > 0.

13. A method for manufacturing a non-volatile memory as claimed in claim 8, wherein the thickness of the oxide semiconductor layer is 2 nm to 200 nm.

14. A method for manufacturing a non-volatile memory as claimed in claim 8, wherein the oxide semiconductor layer is the channel layer of the non-volatile memory.

15. The method of manufacturing a non-volatile memory as claimed in claim 8, further comprising: forming a first spacer wall on the sidewall of the gate stack; and forming a second spacer wall on the sidewall of the erase gate.

16. The method of manufacturing a non-volatile memory as claimed in claim 8, wherein the top surface of the control gate and the top surface of the erase gate are coplanar.