Memory cell and nonvolatile semiconductor storage device
By adopting fin structure and depletion layer technology in the memory cell, the charge injection interference problem caused by high voltage is solved, and the memory cell is miniaturized and high-speed operation is achieved.
Patent Information
- Application Number
- CN202110561992.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2016-08-24
- Filing Date
- 2016-12-07
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2036-12-07
AI Technical Summary
In the conventional nonvolatile semiconductor memory device, high voltage charge storage gate voltage causes undesirable charge injection, causing interference, and miniaturizes the problem in a memory cell arranged in a matrix.
The memory cell with a fin structure forms a fin on a semiconductor substrate and stacks a memory gate insulating film and charge storage layer on the fins, and uses the voltage difference between the fins to generate a quantum tunneling effect for data writing, and at the same time, A depletion layer is formed in the fins to prevent charge injection, miniaturization and interference suppression are achieved.
Effectively prevent charge injection, reduce interference, and miniaturize memory cells by increasing the concentration of impurities in the fin and adjusting the shape of the fin, while reducing the voltage value of the bit line and the source line, and improving the operating speed.
Smart Images

Figure CN113314537B_ABST
Abstract
Description
[0001] This application is a divisional application of the original application number 201680029797.5 (PCT / JP2016 / 086355 entering the Chinese national phase application), the original application date December 7, 2016, and the original invention name "Memory cell, non-volatile semiconductor storage device and method for manufacturing a non-volatile semiconductor storage device". Technical Field
[0002] The present invention relates to a memory unit, a nonvolatile semiconductor storage device and a method for manufacturing the nonvolatile semiconductor storage device. Background Art
[0003] In the prior art, Japanese Patent Publication No. 2011-129816 (Patent Document 1) discloses a memory cell in which a memory gate structure is arranged between two selection gate structures (Patent Document 1, see FIG. 15). In fact, the memory cell includes a drain region connected to a bit line and a source region connected to a source line, and a first selection gate structure, a memory gate structure, and a second selection gate structure are sequentially arranged on a semiconductor substrate between the drain region and the source region. In a memory cell having such a structure, a charge storage layer surrounded by an insulating material is provided on the memory gate structure, and data is written by injecting charge into the charge storage layer, or data is erased by extracting charge from the charge storage layer.
[0004] In fact, in such a memory cell, when injecting charges into the charge storage layer, the voltage is blocked in the second selection gate structure connected to the source line, and at the same time, a low voltage bit voltage from the bit line is applied to the channel layer of the memory gate structure through the first selection gate structure. At this time, in the memory gate structure, a high voltage memory gate voltage is applied to the memory gate, and charges are injected into the charge storage layer through the quantum tunnel effect generated by the voltage difference between the bit voltage and the memory gate voltage.
[0005] In a nonvolatile semiconductor memory device in which a plurality of memory cells having such a structure are arranged in a matrix, a memory gate line for applying a voltage to each memory gate is shared by a plurality of memory cells. Therefore, when a high-voltage charge storage gate voltage is applied to the memory gate line in order to inject charge into a charge storage layer of a predetermined memory cell, a high-voltage charge storage gate voltage is also applied to the memory gates of other memory cells sharing the memory gate line.
[0006] Therefore, in a memory cell that does not inject charge into the charge storage layer, for example, a high voltage bit voltage is applied to the channel layer of the memory gate structure, so that the voltage difference between the memory gate and the channel layer becomes smaller, thereby preventing the injection of charge into the charge storage layer even if a high voltage charge storage gate voltage is applied to the memory gate line.
[0007] Prior art literature
[0008] Patent Literature
[0009] Patent Document 1: Japanese Patent Application Publication No. 2011-129816 Summary of the invention
[0010] Problems to be solved by the invention
[0011] However, in a memory cell that does not inject charge into the charge storage layer, when a high-voltage charge storage gate voltage is applied to the memory gate line, it is not possible to simply prevent the injection of charge into the charge storage layer. At this time, unwanted charges will be injected into the charge storage layer, causing a phenomenon in which the charge storage state of the charge storage layer changes (hereinafter referred to as interference). Therefore, countermeasures need to be taken to prevent this from happening.
[0012] Furthermore, in such a nonvolatile semiconductor memory device, a plurality of memory cells are arranged in a matrix, and therefore even when taking measures against malfunctions caused by interference, it is important to miniaturize the memory cells so as to arrange more memory cells within a limited area.
[0013] Therefore, the present invention has been made in consideration of the above problems, and an object of the present invention is to provide a memory cell, a nonvolatile semiconductor memory device, and a method for manufacturing a nonvolatile semiconductor memory device that can suppress the occurrence of disturbance while achieving miniaturization.
[0014] Technical means to solve the problem
[0015] The memory cell of the present invention for solving the above-mentioned problems is characterized in that it comprises: a semiconductor substrate covered by an insulating layer; a fin portion formed on the semiconductor substrate in a manner protruding from the insulating layer; a memory gate structure stacked with a lower memory gate insulating film, a charge storage layer, an upper memory gate insulating film and a memory gate, and formed on the insulating layer in a manner straddling the fin portion; a first selection gate structure, a first selection gate is provided on the first selection gate insulating film, and a side wall partition formed on one side wall of the memory gate structure is formed on the insulating layer in a manner straddling the fin portion; a second selection gate structure, a second selection gate is provided on the second selection gate insulating film, and a second side wall partition formed on the other side wall of the memory gate structure is formed on the insulating layer in a manner straddling the fin portion; and a drain region is formed on the surface of the fin portion adjacent to the first selection gate structure in a manner parallel to the first selection gate. The fin is provided in an insulated manner and electrically connected to a bit line; the source region is provided on the surface of the fin adjacent to the second selection gate structure in an insulated manner from the second selection gate and electrically connected to an active line, the first selection gate structure, the memory gate structure and the second selection gate structure are provided between the drain region and the source region, when the protruding height of the electrode of the fin, which is the distance from the upper surface of the fin to the bottom surface of the memory gate on the insulating layer, is represented by Hfin, and the width of the fin in the direction across the fin by the first selection gate structure and the second selection gate structure is represented by Wfin, Hfin>Wfin, when writing is selected, charges are injected into the charge storage layer through a quantum tunneling effect based on a voltage difference between the memory gate and the fin, and when writing is not selected, charge injection into the charge storage layer is prevented by a depletion layer formed in the fin.
[0016] In addition, the nonvolatile semiconductor memory device according to the present invention is characterized in that the memory cells connected to the memory gate lines are arranged in a matrix, the memory cells are the above-mentioned memory cells, and the plurality of memory cells arranged in a matrix share the memory gate lines.
[0017] Effects of the Invention
[0018] According to the memory cell, non-volatile semiconductor memory device and method for manufacturing a non-volatile semiconductor memory device of the present invention, the following memory cell can be provided: when preventing charge injection into the charge storage layer, the electrical connection between the fin surrounded by the memory gate structure and the bit line is blocked by the first selection gate structure, and the electrical connection between the fin surrounded by the memory gate structure and the source line is blocked by the second selection gate structure, thereby forming a depletion layer on the surface of the fin, and through the depletion layer, the potential difference between the memory gate and the fin is reduced.
[0019] Therefore, in the present invention, even if the impurity concentration in the fin is increased to bring the drain region and the source region closer together on the surface of the fin to achieve miniaturization, the shape of the fin is selected in such a way that a depletion layer is formed in the entire fin, thereby making it possible to reduce the potential difference between the memory gate and the fin, and further reduce the electric field of the depletion layer, thereby suppressing the occurrence of interference.
[0020] In addition, in the memory unit of the present invention, the gate widths of the first selection gate structure, the memory gate structure, and the second selection gate structure can be replaced by the height of the fin. Therefore, although the height is increased by an amount equivalent to the height of the fin, the formation area of the first selection gate structure, the memory gate structure, and the second selection gate structure in the direction of the gate width can be reduced accordingly, thereby achieving miniaturization accordingly.
[0021] Incidentally, in the memory cell of the present invention, when preventing the injection of charges into the charge storage layer, it is sufficient to apply a voltage to the bit line and the source line that can block the electrical connection between the fin portion surrounded by the memory gate structure and the bit line (source line). Therefore, in the memory cell of the present invention, the voltage values of the bit line and the source line can be reduced without being restricted by the charge storage gate voltage applied to the memory gate, and the thickness of each of the first selection gate insulating film of the first selection gate structure and the second selection gate insulating film of the second selection gate structure can be reduced accordingly, so that high-speed operation can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS
[0022] Figure 1 It is a perspective view showing the structure (1) of a memory cell.
[0023] Figure 2 is a schematic diagram showing a planar layout of a nonvolatile semiconductor memory device.
[0024] Figure 3A It is shown Figure 2 Schematic diagram of the cross-sectional structure of the A-A' portion, Figure 3B It is shown Figure 2 Schematic diagram of the cross-sectional structure of the BB' portion.
[0025] Figure 4A It is shown Figure 2 A schematic diagram of the cross-sectional structure of the memory cell at the CC' portion, Figure 4B It is shown Figure 2 Schematic diagram of the cross-sectional structure of the memory cell at the D-D' portion.
[0026] Figure 5 1 is a schematic diagram showing a circuit structure of a nonvolatile semiconductor memory device in which a plurality of memory cells are arranged in a matrix.
[0027] Figure 6 This is a table showing an example of voltage values at various locations during a data writing operation, a data reading operation, and a data erasing operation.
[0028] Figure 7 It is a schematic diagram for explaining the potential written into the non-selected memory cell.
[0029] Figure 8 Schematic diagram showing a manufacturing process (1) of a nonvolatile semiconductor memory device.
[0030] Fig. 9A is a schematic diagram showing a manufacturing process (2) of a nonvolatile semiconductor memory device, Fig. 9B It is shown Fig. 9A Schematic diagram of the cross-sectional structure of the E-E' portion, Fig. 9C is a schematic diagram showing a manufacturing process (3) of a nonvolatile semiconductor memory device, Fig.9D It is shown Fig. 9C Schematic diagram of the cross-sectional structure of the E-E' portion.
[0031] Fig. 10A It is shown Fig.9D A schematic diagram of a manufacturing process (1) of a nonvolatile semiconductor memory device at a position shown in FIG. Fig. 10B It is shown Fig.9D A schematic diagram of a manufacturing process (2) of a nonvolatile semiconductor memory device at a position shown in FIG. Fig. 10C yes Fig.9D Schematic diagram of the manufacturing process (3) of the nonvolatile semiconductor memory device at the position shown.
[0032] Fig.11A is a schematic diagram showing a manufacturing process (4) of a nonvolatile semiconductor memory device, Fig. 11B It is shown Fig.11A Schematic diagram of the cross-sectional structure of the E-E' portion.
[0033] Fig. 12A It is shown Figure 2 A schematic diagram of a cross-sectional structure of a memory cell according to another embodiment of the present invention is shown in section AA'. Fig. 12BIt is shown Figure 2 Portion BB' is a schematic diagram of a cross-sectional structure of a memory cell according to another embodiment.
[0034] Fig.13 It is a perspective view showing the structure (2) of a memory cell.
[0035] Fig.14A yes Fig.13 Schematic diagram of the cross-sectional structure of the A-A' portion, Fig. 14B yes Fig.13 Schematic diagram of the cross-sectional structure of the BB' portion.
[0036] Fig.15A It is shown Fig.13 A schematic diagram of the cross-sectional structure of the memory cell at the CC' portion, Fig. 15B It is shown Fig.13 Schematic diagram of the cross-sectional structure of the memory cell at the D-D' portion.
[0037] Fig.16 It is a perspective view showing the structure (3) of the memory cell.
[0038] Fig.17A yes Fig.16 Schematic diagram of the cross-sectional structure of the A-A' portion, Fig. 17B yes Fig.16 Schematic diagram of the cross-sectional structure of the BB' portion.
[0039] Fig.18A It is shown Fig.16 A schematic diagram of the cross-sectional structure of the memory cell at the CC' portion, Fig.18B It is shown Fig.16 Schematic diagram of the cross-sectional structure of the memory cell at the D-D' portion.
[0040] Fig.19 It is a perspective view showing the structure (4) of the memory cell.
[0041] Fig. 20A yes Fig.19 Schematic diagram of the cross-sectional structure of the A-A' portion, Fig. 20B yes Fig.19 Schematic diagram of the cross-sectional structure of the BB' portion.
[0042] Fig.21A It is shown Fig.19 A schematic diagram of the cross-sectional structure of the memory cell at the CC' portion, Fig.21B It is shown Fig.19 Schematic diagram of the cross-sectional structure of the memory cell at the D-D' portion.
[0043] Specific embodiments for implementing the invention
[0044] Hereinafter, embodiments of the present invention will be described in the following order.
[0045] <1. Overview of Memory Cell Having Fin Structure>
[0046] <2. Detailed structure of memory cell>
[0047] <3. Circuit Structure of Nonvolatile Semiconductor Memory Device>
[0048] <4. Voltages of Nonvolatile Semiconductor Memory Devices During Various Operations>
[0049] 4-1. Data writing operation
[0050] 4-2. Non-write data operation
[0051] 4-3. Data reading operation
[0052] 4-4. Data erasure operation
[0053] <5. Regarding the potential of the high voltage charge storage gate voltage applied to the memory gate for writing non-selected memory cells>
[0054] <6. Method for manufacturing nonvolatile semiconductor memory device>
[0055] <7. Function and Effect>
[0056] <8. Other Implementation Methods>
[0057] 8-1. Structure of a memory cell when the memory gate, the first selection gate, and the second selection gate are formed of a conductive material other than a metal material
[0058] 8-2. Structure of a memory cell having a cap insulating film provided on the upper surface of the fin
[0059] 8-3. Others
[0060] <9. Memory Cell in Which the Film Thickness of the Insulating Layer in the Region Where the First Select Gate Structure and the Second Select Gate Structure Are Provided Is Reduced>
[0061] 9-1. Structure of memory cell
[0062] 9-2. Action and Effect
[0063] <10. Memory cell having fin on insulating layer>
[0064] 10-1. Structure of memory cell
[0065] 10-2. Function and effect
[0066] 10-3. Fin according to other embodiments
[0067] (1) Overview of Memory Cell Having Fin Structure
[0068] Figure 1 FIG. 1 is a perspective view schematically showing a fin structure of a memory cell MC provided in the nonvolatile semiconductor memory device 1. Figure 1 The stereoscopic diagram shown briefly explains that the memory cell MC has a fin-type FET (Field Effect Transistor) structure. In this case, the memory cell MC has a semiconductor substrate S1 formed of a semiconductor material such as silicon, and a fin portion S2 formed integrally with the semiconductor substrate S1 and formed of the same semiconductor material as the semiconductor substrate S1, and the fin portion S2 is formed to protrude from the semiconductor substrate S1.
[0069] In fact, the memory cell MC is configured such that the surface of the semiconductor substrate S1 formed in a plate shape is covered by an insulating layer IS composed of an insulating material, and the fin S2 is formed on the semiconductor substrate S1 in a manner extending in the vertical direction z, and a part of the fin S2 protrudes from the surface of the insulating layer IS. The fin S2 is formed in a rectangular parallelepiped shape, and the length direction extends in the y direction along the surface of the semiconductor substrate S1. In addition, in this embodiment, for example, P-type impurities are injected into the semiconductor substrate S1 and the fin S2.
[0070] In addition, in this embodiment, in the memory cell MC, a memory gate structure 2 for forming an N-type memory transistor MT in the fin S2, a first selection gate structure 3 for forming a first selection transistor T1 of an N-type MOS in the fin S2, and a second selection gate structure 4 for forming a second selection transistor T2 of the same N-type MOS in the fin S2 are formed on the insulating layer IS in a manner that spans the fin S2.
[0071] In addition, the detailed structure of the memory gate structure 2, the first selection gate structure 3 and the second selection gate structure 4 will be described in Figures 3 and 4 described later, so their description is omitted here, and the structure of the memory gate structure 2, the first selection gate structure 3 and the second selection gate structure 4 across the fin S2 is mainly described.
[0072] In this case, the memory gate structure 2, the first selection gate structure 3, and the second selection gate structure 4 are extended on the insulating layer IS in a manner that their length directions intersect with the y direction in which the fin S2 extends, and are arranged in a manner that the surface of the fin S2 is covered in a "U" shape. As a result, the fin S2 has a structure in which the surface (the opposite side surface and the upper surface) exposed on the surface of the insulating layer IS is covered by the memory gate structure 2, the first selection gate structure 3, and the second selection gate structure 4.
[0073] Furthermore, in this embodiment, the memory gate structure 2, the first selection gate structure 3, and the second selection gate structure 4 are arranged in parallel and are arranged so as to be perpendicular to each other in the x direction across the fin S2, the y direction in which the fin S2 is extended, and the z direction perpendicular to the fin S2. In addition, the memory gate structure 2 is insulated from the first selection gate structure 3 by a side wall spacer 6 formed along one side wall, and is insulated from the second selection gate structure 4 by another side wall spacer 7 formed along another side wall.
[0074] In the fin S2, a drain region 12a is formed on a surface adjacent to the first selection gate structure 3 in a manner insulated from the first selection gate DG of the first selection gate structure 3, and a bit line (not shown) is connected to the drain region 12a. In addition, in the fin S2, a source region 12b is formed on a surface adjacent to the second selection gate structure 4 in a manner insulated from the second selection gate SG of the second selection gate structure 4, and a source line (not shown) is connected to the source region 12b.
[0075] (2) Detailed structure of memory cell
[0076] Next, use Figure 2 , Figure 3A , Figure 3B , Figure 4A and Figure 4B , the detailed structure of the memory cell MC is described, wherein, Figure 2 4 shows a planar layout of a nonvolatile semiconductor memory device 1 in which, for example, four memory cells MC are arranged side by side in the x direction. Figure 3A Show Figure 1 and Figure 2 The cross-sectional structure of the A-A' section, Figure 3B Show Figure 1 and Figure 2 The cross-sectional structure of the B-B' part, Figure 4A Show Figure 1 and Figure 2 The cross-sectional structure of the C-C' part, Figure 4B Show Figure 1 and Figure 2 The cross-sectional structure of the D-D' portion.
[0077] and, Figure 2 The structure of the fin S2, the memory gate structure 2, the first selection gate structure 3 and the second selection gate structure 4 are mainly shown. Figure 2 Omitted in Figure 1 The structures shown include a sidewall spacer 6 between the memory gate structure 2 and the first select gate structure 3 , a sidewall spacer 7 between the memory gate structure 2 and the second select gate structure 4 , and the like.
[0078] In this case, if Figure 2 As shown, in the nonvolatile semiconductor memory device 1, a plurality of fins S2 are arranged side by side in the y direction, and a memory cell MC is formed in each fin S2. In the nonvolatile semiconductor memory device 1, a memory gate structure 2, a first selection gate structure 3, and a second selection gate structure 4 are extended side by side in a manner intersecting the y direction in which each fin S2 is extended, and a memory cell MC having a memory transistor MT, a first selection transistor T1, and a second selection transistor T2 is formed in a region where the memory gate structure 2, the first selection gate structure 3, and the second selection gate structure 4 cross each fin S2.
[0079] In the non-volatile semiconductor memory device 1, a memory gate connector MGC connected to a memory gate line (not shown) is provided at a predetermined position on the upper surface of the memory gate MG of the memory gate structure 2, and the memory gate voltage applied to the memory gate line can be applied to the memory gate MG shared by multiple memory cells MC through the memory gate connector MGC.
[0080] In addition, in the non-volatile semiconductor memory device 1, a first selection gate connector DGC connected to a first selection gate line (not shown) is provided at a predetermined position on the upper surface of the first selection gate DG of the first selection gate structure 3, and the voltage applied to the first selection gate line can be applied to the first selection gate DG shared by multiple memory cells MC through the first selection gate connector DGC.
[0081] Furthermore, in the non-volatile semiconductor memory device 1, a second selection gate connector SGC connected to a second selection gate line (not shown) is provided at a predetermined position on the upper surface of the second selection gate SG of the second selection gate structure 4, and the voltage applied to the second selection gate line can be applied to the second selection gate SG shared by multiple memory cells MC through the second selection gate connector SGC.
[0082] A drain region 12a is formed on the surface of the fin S2 adjacent to the first selection gate structure 3, and a bit connector BC connected to a bit line (not shown) is provided in the drain region 12a. Thus, in the memory cell MC, a bit voltage applied to the bit line is applied to the drain region 12a through the bit connector BC.
[0083] On the other hand, a source region 12b is formed on the surface of the fin S2 adjacent to the second selection gate structure 4, and a source connector SC connected to a source line (not shown) is provided in the source region 12b. Thus, in the memory cell MC, a source voltage applied to the source line is applied to the source region 12b through the source connector SC.
[0084] In fact, in the memory cell MC, as shown Figure 1 and Figure 2 The cross-sectional structure of the A-A' section Figure 3A As shown, a sidewall portion 22a made of an insulating material such as silicon oxide (SiO, SiO2) is formed along the sidewall of the first selection gate structure 3, and a drain region 12a is formed on the surface of the fin S2 adjacent to the sidewall portion 22a. The drain region 12a is insulated from the first selection gate DG of the first selection gate structure 3 by the sidewall portion 22a formed between the first selection gate structure 3 and the drain region 12a.
[0085] Furthermore, in the memory cell MC, a sidewall portion 22b made of an insulating material such as silicon oxide (SiO, SiO2) is formed along the sidewall of the second selection gate structure 4, and a source region 12b is formed on the surface of the fin S2 adjacent to the sidewall portion 22b. The source region 12b is insulated from the second selection gate SG of the second selection gate structure 4 by the sidewall portion 22b formed between the second selection gate structure 4 and the source region 12b.
[0086] Furthermore, in this embodiment, the drain region 12a and the source region 12b formed at a predetermined interval on the surface of the fin S2 are composed of a semiconductor material such as SiGe, and are selectively formed on the surface of the fin S2 by an epitaxial growth method to have a predetermined film thickness.
[0087] In the memory gate structure 2, a lower memory gate insulating film 13 composed of an insulating material such as silicon oxide (SiO, SiO2) is sandwiched on the fin S2 between the drain region 12a and the source region 12b, and a charge storage layer EC composed of, for example, silicon nitride (Si3N4), silicon oxynitride (SiON), aluminum oxide (Al2O3), hafnium dioxide (HfO2), etc. is provided, and further, an upper memory gate insulating film 14 composed of an insulating material different from the lower memory gate insulating film 13 (for example, a high-k material such as hafnium oxide (HfO2), hafnium silicate nitride (HfSiON)) is sandwiched on the charge storage layer EC and has a memory gate MG. Thus, the memory gate structure 2 has a structure in which the charge storage layer EC is insulated from the fin S2 and the memory gate MG by the lower memory gate insulating film 13 and the upper memory gate insulating film 14.
[0088] The memory gate MG is formed of a metal material such as aluminum (Al), titanium aluminum (TiAl), tantalum carbide (TaC), tantalum silicon nitride (TaSiN), etc., and its upper surface is flattened by a flattening process such as CMP performed in the manufacturing process. In addition, a memory gate line ML is connected to the memory gate MG, and a predetermined voltage can be applied from the memory gate line ML.
[0089] In the memory gate structure 2, a sidewall spacer 6 made of an insulating material is formed along one side wall, and the first selection gate structure 3 is adjacent to the sidewall spacer 6. In this embodiment, in the memory gate structure 2, a wall-shaped memory gate sidewall insulating film 15a formed integrally with one end of the upper memory gate insulating film 14 is provided along one side wall of the memory gate MG, and the sidewall spacer 6 is formed along each side wall of the memory gate sidewall insulating film 15a, the upper memory gate insulating film 14, the charge storage layer EC, and the lower memory gate insulating film 13. In addition, the memory gate sidewall insulating film 15a formed in the memory gate structure 2 can be formed of the same insulating material (for example, high-k material) as the upper memory gate insulating film 14, and can be formed by the same manufacturing process as the upper memory gate insulating film 14.
[0090] The sidewall spacer 6 formed between the memory gate structure 2 and the first selection gate structure 3 is formed to a predetermined film thickness, and together with the memory gate sidewall insulating film 15a in the memory gate structure 2 and the first selection gate sidewall insulating film 19a in the first selection gate structure 3 to be described later, insulates the memory gate MG and the first selection gate DG.
[0091] In this case, the sidewall spacer 6 is formed of an insulating material such as silicon oxide (SiO, SiO2) that is different from the insulating material (for example, high-k material) of the memory gate sidewall insulating film 15a in the memory gate structure 2 and the first selection gate sidewall insulating film 19a in the first selection gate structure 3, and its upper surface is flattened by a flattening treatment such as CMP performed in the manufacturing process.
[0092] Here, when the distance between the memory gate MG and the first selection gate DG is less than 5 nm, when a predetermined voltage is applied to the memory gate MG and the first selection gate DG, poor voltage resistance of the memory gate sidewall insulating film 15a, the sidewall spacer 6 and the first selection gate sidewall insulating film 19a may occur.
[0093] On the other hand, when the distance between the memory gate MG and the first selection gate DG exceeds 40nm, the impedance in the fin S2 (for example, a region 50nm deep from the surface (surface region)) between the memory gate MG and the first selection gate DG increases, making it difficult for a read current to flow between the memory gate structure 2 and the first selection gate structure 3 when reading data.
[0094] Therefore, in this embodiment, the distance between the memory gate MG and the first selection gate DG is preferably set to be greater than 5 nm and less than 40 nm, and therefore the total film thickness of the memory gate sidewall insulating film 15a, the sidewall spacer 6 and the first selection gate sidewall insulating film 19a is also preferably set to be greater than 5 nm and less than 40 nm.
[0095] Furthermore, the sidewall spacer 6 is preferably formed of an insulating material having a lower dielectric constant than the memory gate sidewall insulating film 15a and the first select gate sidewall insulating film 19a. In this case, the capacitance between the first select gate structure 3 and the memory gate structure 2 is reduced, thereby improving the access speed.
[0096] In the first selection gate structure 3, a lower first selection gate insulating film 17a made of an insulating material such as silicon oxide (SiO, SiO2) is formed on the fin portion S2 between the side wall spacer 6 and the side wall portion 22a, and an upper first selection gate insulating film 18a made of an insulating material different from the first selection gate insulating film 17a (for example, a high-k material) is formed on the lower first selection gate insulating film 17a.
[0097] In addition, in the first selection gate structure 3, the first selection gate sidewall insulating film 19a formed in a wall shape along the sidewall spacer 6 and the sidewall sidewall insulating film 20a formed in a wall shape along the sidewall of the sidewall portion 22a are formed integrally with the upper first selection gate insulating film 18a. Also, the first selection gate sidewall insulating film 19a and the sidewall sidewall insulating film 20a are also formed of an insulating material (for example, a high-k material) different from the insulating material of the lower first selection gate insulating film 17a and the sidewall spacer 6, as with the upper first selection gate insulating film 18a, and are formed together with the upper first selection gate insulating film 18a through the same manufacturing process. Here, the total film thickness of the lower first selection gate insulating film 17a and the upper first selection gate insulating film 18a is formed to be 9 nm or less, preferably 3 nm or less.
[0098] In addition, in the first selection gate structure 3, the first selection gate DG is formed on the upper first selection gate insulating film 18a, and the first selection gate sidewall insulating film 19a and the sidewall insulating film 20a are formed along the sidewalls of the first selection gate DG.
[0099] The first selection gate DG is formed of the same metal material as the memory gate MG (for example, aluminum (Al), titanium aluminum (TiAl), tantalum carbide (TaC), tantalum silicon nitride (TaSiN), etc.), and its upper surface is flattened by a flattening process such as CMP performed in the manufacturing process. In addition, a first selection gate line DL is connected to the first selection gate DG, and a predetermined voltage can be applied from the first selection gate line DL.
[0100] On the other hand, a wall-shaped sidewall spacer 7 made of an insulating material is also formed on the other sidewall of the memory gate structure 2, and the second selection gate structure 4 is adjacent to the sidewall spacer 7. In this embodiment, in the memory gate structure 2, a wall-shaped memory gate sidewall insulating film 15b formed integrally with the other end of the upper memory gate insulating film 14 is also provided on the other sidewall of the memory gate MG, and another sidewall spacer 7 is formed along the sidewalls of the memory gate sidewall insulating film 15b, the upper memory gate insulating film 14, the charge storage layer EC, and the lower memory gate insulating film 13. In addition, the memory gate sidewall insulating film 15b formed in the memory gate structure 2 is formed of the same insulating material (for example, high-k material) as the upper memory gate insulating film 14 and the one memory gate sidewall insulating film 15a, and can be formed by the same manufacturing process as the upper memory gate insulating film 14 and the one memory gate sidewall insulating film 15a.
[0101] The sidewall spacer 7 formed between the memory gate structure 2 and the second selection gate structure 4 is formed to have the same film thickness as the sidewall spacer 6, and together with the memory gate sidewall insulating film 15b in the memory gate structure 2 and the second selection gate sidewall insulating film 19b in the second selection gate structure 4 to be described later, insulates the memory gate MG and the second selection gate SG.
[0102] In this case, the side wall spacer 7 is formed of the same insulating material as the side wall spacer 6, such as silicon oxide (SiO, SiO2), and its upper surface is flattened by a flattening treatment such as CMP performed in the manufacturing process.
[0103] Here, as with the above-mentioned problem between the memory gate MG and the first selection gate DG, between the memory gate MG and the second selection gate SG, there may also be a problem of poor voltage resistance of the memory gate sidewall insulating film 15b, the sidewall spacer 7, and the second selection gate sidewall insulating film 19b, as well as an undesirable phenomenon of a decrease in the reading current between the memory gate structure 2 and the second selection gate structure 4.
[0104] Therefore, in this case, the total film thickness of the memory gate sidewall insulating film 15b, the sidewall spacer 7, and the second selection gate sidewall insulating film 19b is preferably set to be greater than 5nm and less than 40nm. In addition, it is preferred that the other sidewall spacer 7 is also formed of an insulating material having a dielectric constant smaller than that of the memory gate sidewall insulating film 15b and the second selection gate sidewall insulating film 19b. In this case, the capacitance between the second selection gate structure 4 and the memory gate structure 2 becomes smaller, thereby being able to improve the access speed.
[0105] In the second selection gate structure 4, a lower second selection gate insulating film 17b made of an insulating material such as silicon oxide (SiO, SiO2) is formed on the fin portion S2 between the side wall spacer 7 and the side wall portion 22b, and an upper second selection gate insulating film 18b made of an insulating material different from the second selection gate insulating film 17b (for example, a high-k material) is formed on the lower second selection gate insulating film 17b.
[0106] In addition, in the second selection gate structure 4, the second selection gate sidewall insulating film 19b formed in a wall shape along the sidewall spacer 7 and the sidewall sidewall insulating film 20b formed in a wall shape along the sidewall of the sidewall portion 22b are formed integrally with the upper second selection gate insulating film 18b. Also, the second selection gate sidewall insulating film 19b and the sidewall sidewall insulating film 20b are also formed of an insulating material (for example, a high-k material) different from the insulating material of the lower second selection gate insulating film 17b and the sidewall spacer 7, as with the upper second selection gate insulating film 18b, and are formed together with the upper second selection gate insulating film 18b through the same manufacturing process. Here, the total film thickness of the lower second selection gate insulating film 17b and the upper second selection gate insulating film 18b is formed to be 9 nm or less, preferably 3 nm or less.
[0107] In addition, in the second selection gate structure 4, a second selection gate SG is formed on the upper second selection gate insulating film 18b, and a second selection gate sidewall insulating film 19b and a sidewall insulating film 20b are formed along the sidewalls of the second selection gate SG.
[0108] The second selection gate SG is formed of the same metal material as the memory gate MG (for example, aluminum (Al), titanium aluminum (TiAl), tantalum carbide (TaC), tantalum silicon nitride (TaSiN), etc.), and its upper surface is flattened by a flattening process such as CMP performed in the manufacturing process. In addition, a second selection gate line SGL is connected to the second selection gate SG, and a predetermined voltage is applied from the second selection gate line SGL.
[0109] Here, in the memory cell MC, the upper surfaces of the memory gate structure 2, the first selection gate structure 3, the second selection gate structure 4 and the sidewall spacers 6 and 7 are flattened and aligned to the same height position, so no protruding area is formed, thereby facilitating the processing of the upper layer.
[0110] Incidentally, in the memory cell MC, since the memory gate MG, the first selection gate DG, and the second selection gate SG are formed of a predetermined metal material, it is possible to prevent the memory gate MG, the first selection gate DG, and the second selection gate SG from being depleted.
[0111] Furthermore, in the non-volatile semiconductor memory device 1, the periphery and upper surface of the memory cell MC, the insulating layer IS on the semiconductor substrate S1, the fin S2 protruding from the insulating layer IS, etc. are covered by an interlayer insulating layer 25 composed of insulating materials such as silicon oxide (SiO, SiO2) and an upper interlayer insulating layer (not shown).
[0112] Next, show Figure 1 and Figure 2The cross-sectional structure of the B-B' section Figure 3B For explanation. Figure 3B As shown, in the position where the fin S2 is not formed, Figure 3A The memory gate structure 2, the first selection gate structure 3, the second selection gate structure 4, the sidewall spacers 6, 7 and the sidewalls 22a, 22b shown in FIG. are formed on the insulating layer IS covering the semiconductor substrate S1. The upper surfaces of the first selection gate structure 3, the memory gate structure 2 and the second selection gate structure 4 on the insulating layer IS are flattened to be the same as Figure 3A The first select gate structure 3 , the memory gate structure 2 , and the second select gate structure 4 on the fin S2 are at the same height.
[0113] Here, at a position where the fin S2 is not formed, the surface of the insulating layer IS is located at a position lower than the surface of the fin S2, and accordingly, the first selection gate structure 3, the memory gate structure 2, the second selection gate structure 4, the sidewall spacers 6, 7 and the sidewall portions 22a, 22b are formed longitudinally compared to the position having the fin S2, thereby the first selection gate structure 3, the memory gate structure 2, the second selection gate structure 4, the sidewall spacers 6, 7 and the sidewall portions 22a, 22b become the same height as the position having the fin S2.
[0114] In addition, in the memory cell MC, the distance H1 from the surface of the semiconductor substrate S1 to the lower surface of the memory gate MG is greater than the distance H2 from the surface of the semiconductor substrate S1 to the lower surfaces of the first selection gate DG and the second selection gate SG, for example, by more than the film thickness of the charge storage layer EC, and the lower surface positions of the first selection gate DG and the second selection gate SG are configured at positions lower than the lower surface position of the memory gate MG.
[0115] Thus, in the memory cell MC, the first selection gate DG and the second selection gate SG are formed at the lower surface position lower than the lower surface position of the memory gate MG, so that the first selection transistor T1 and the second selection transistor T2 in the side area of the fin S2 can be fully controlled to be turned on and off. Therefore, in the memory cell MC, it is possible to prevent malfunctions from occurring when writing or reading data.
[0116] Here, as a comparison with the memory cell MC, the following memory cell (hereinafter referred to as a comparative example) is described below. This memory cell is different from the memory cell MC, and at a position where the fin S2 is not formed, the distance H1 from the surface of the semiconductor substrate S1 to the lower surface of the memory gate MG is smaller than the distance H2 from the surface of the semiconductor substrate S1 to the lower surfaces of the first selection gate DG and the second selection gate SG, and the lower surface positions of the first selection gate DG and the second selection gate SG are arranged at a position higher than the lower surface position of the memory gate MG.
[0117] In the comparative example, the lower surface positions of the first selection gate DG and the second selection gate SG are arranged at a position higher than the lower surface position of the memory gate MG, so the first selection gate DG and the second selection gate SG cannot fully control the conduction and cutoff of the first selection transistor T1 and the second selection transistor T2 in the side area of the fin S2. In particular, in this comparative example, even if a gate cutoff voltage is applied to the first selection gate DG and the second selection gate SG, the side area of the fin S2 not covered by the first selection gate DG and the second selection gate SG cannot be made non-conductive. When the information written to the memory transistor MT is a low threshold voltage Vth, a leakage current flows between the drain area 12a and the source area 12b along the side area of the fin S2, and a malfunction may occur.
[0118] Here, as shown Figure 1 and Figure 2 The cross-sectional structure of the C-C' part Figure 4A As shown, the memory gate structure 2 is formed on the insulating layer IS in a manner spanning the fin S2, so the lower memory gate insulating film 13, the charge storage layer EC, the upper memory gate insulating film 14 and the memory gate MG can be arranged in a manner surrounding the surface of the fin S2. Thus, the memory gate structure 2 can form a channel layer of the memory transistor MT along the surface of the fin S2 surrounded by the memory gate structure 2.
[0119] As described above, in the memory gate structure 2, the gate width of the memory transistor MT is the distance along the surface of the fin S2 surrounded by the memory gate structure 2, so a part of the gate width is replaced by the height along the side of the fin S2, so that the formation area of the memory transistor MT can be narrowed accordingly.
[0120] In addition, as shown Figure 1 and Figure 2 The cross-sectional structure of the D-D' part Figure 4BAs shown, the first selection gate structure 3 is also formed on the insulating layer IS in a manner spanning the fin S2, so the lower first selection gate insulating film 17a, the upper first selection gate insulating film 18a and the first selection gate DG are provided in a manner surrounding the surface of the fin S2. Thus, in the first selection gate structure 3, the channel layer of the first selection transistor T1 can also be formed along the surface of the fin S2 surrounded by the first selection gate structure 3.
[0121] As described above, in the first selection gate structure 3, the gate width of the first selection transistor T1 is also the distance along the surface of the fin S2 surrounded by the first selection gate structure 3, so a part of the gate width is replaced by the height along the side of the fin S2, and the formation area of the first selection transistor T1 can be narrowed accordingly. Figure 4B The first selection gate structure 3 shown has the same structure, so the description thereof is omitted here.
[0122] Here, in this embodiment, if Figure 4A and Figure 4B As shown, the memory cell MC is formed such that when the distance from the upper surface of the fin S2 to the bottom surface of the memory gate MG on the insulating layer IS (the protruding height within the electrode of the fin S2) is represented by Hfin, and when the width of the fin S2 in the x direction of the fin S across the first selection gate structure 3, the memory gate structure 2 and the second selection gate structure 4 is represented by Wfin, the relationship Hfin>Wfin is satisfied.
[0123] In addition, as shown in Figure 3, the memory cell MC is formed so that when the gate length of the first selection gate DG in the y direction extending the fin S2 is represented by L1 and the gate length of the second selection gate SG is represented by L2, the relationship between them and the width Wfin of the fin S2 satisfies L1≦1.5·Wfin, L2≦1.5·Wfin.
[0124] (3) Circuit Structure of Nonvolatile Semiconductor Memory Device
[0125] Next, the circuit structure of the nonvolatile semiconductor storage device in which the memory cells MC are arranged in a matrix will be described. Figure 5As described above, the nonvolatile semiconductor memory device 1 has a structure in which a plurality of memory cells MC11, MC12, ..., MC1n, MC21, MC22, ..., MC2n, MCm1, MCm2, ..., MCmn having the same structure as the above-mentioned memory cell M are arranged in a matrix, and these memory cells MC11, MC12, ..., MC1n, MC21, MC22, ..., MC2n, MCm1, MCm2, ..., MCmn are formed on the same semiconductor substrate S1. In these memory cells MC11, MC12, ..., MC1n, MC21, MC22, ..., MC2n, MCm1, MCm2, ..., MCmn, a predetermined substrate voltage can be uniformly applied to the common semiconductor substrate S1 by a substrate voltage application circuit (not shown). Furthermore, on the semiconductor substrate S1, although a substrate voltage application circuit (not shown) is integrally formed as follows: Figure 1 The fin S2 is shown, but the fin S2 is omitted here.
[0126] In the nonvolatile semiconductor memory device 1, the memory cells MC11, MC12, ..., MC1n, MC21, MC22, ..., MC2n, MCm1, MCm2, ..., MCmn) arranged in one direction (in this case, the column direction) respectively share a bit line BL1 (BL2, BLn), and a predetermined bit voltage is applied to each bit line BL1, BL2, ..., BLn through a bit line voltage application circuit (not shown). In addition, in the non-volatile semiconductor memory device 1, memory cells MC11, MC12, ..., MC1n (MC21, MC22, ..., MC2n, MCm1, MCm2, ..., MCmn) arranged in another direction intersecting one direction (in this case, the row direction) respectively share a first selection gate line DL1 (DL2, DLm), and a predetermined first selection gate voltage is uniformly applied to each first selection gate line DL1, DL2, ..., DLm through a first selection gate voltage application circuit (not shown).
[0127] Further, in this embodiment, in the non-volatile semiconductor memory device 1, a memory gate line ML, a second selection gate line SGL, and a source line SL are shared by all memory cells MC11, MC12, ..., MC1n, MC21, MC22, ..., MC2n, MCm1, MCm2, ..., MCmn located on a semiconductor substrate S1, a predetermined memory gate voltage is applied to the memory gate line ML through a memory gate voltage application circuit (not shown), a second selection gate voltage is applied to the second selection gate line SGL through a second selection gate voltage application circuit (not shown), and a predetermined source voltage is applied to the source line SL through a source line voltage application circuit (not shown).
[0128] Furthermore, in this embodiment, a case is described in which one memory gate line ML, one second selection gate line SGL, and one source line SL are shared by all memory cells MC11, MC12, ..., MC1n, MC21, MC22, ..., MC2n, MCm1, MCm2, ..., MCmn, but the present invention is not limited to this. Memory cells MC11, MC12, ..., MC1n (MC21, MC22, ..., MC2n, MCm1, MCm2, ..., MCmn) arranged in another direction (row direction) may respectively share the memory gate line, the second selection gate line, and the source line.
[0129] Incidentally, in the memory cell MC11, the memory gate MG of the memory gate structure 2 is connected to the memory gate line ML, the first selection gate DG of the first selection gate structure 3 is connected to the first selection gate line DL1, and the second selection gate SG of the second selection gate structure 4 is connected to the second selection gate line SGL. In addition, in the memory cell MC11, one end (drain region) of the first selection transistor T1 formed by the first selection gate structure 3 is connected to the bit line BL1, and one end (source region) of the second selection transistor T2 formed by the second selection gate structure 4 is connected to the source line SL.
[0130] (4) Voltages during various operations of the nonvolatile semiconductor memory device
[0131] Next, various operations of the nonvolatile semiconductor memory device 1 will be described. Figure 6 It is shown Figure 5A table is shown of an example of voltage values at various parts of the non-volatile semiconductor memory device 1 during a writing operation ("Prog") of injecting charge into the charge storage layer EC of a memory cell MCxy (where x is any one of 1, 2, ..., m, and y is any one of 1, 2, ..., n), a reading operation ("Read") of detecting whether there is charge stored in the charge storage layer EC of the memory cell MCxy, and an erasing operation ("Erase") of extracting charge from the charge storage layer EC of the memory cell MCxy.
[0132] and, Figure 6 In the "Write" column of , the column of the memory cell MCxy configured to inject charge into the charge storage layer EC is represented as a "selected column", and the row of the memory cell MCxy configured to inject charge into the charge storage layer EC is represented as a "selected row". In addition, Figure 6 In the "Write" column, a column that is configured with only memory cells MCxy that do not inject charge into the charge storage layer EC is represented as a "non-selected column", and a row that is configured with only memory cells MCxy that do not inject charge into the charge storage layer EC is represented as a "non-selected row". In addition, Figure 6 "DLx" in the figure represents the first selection gate lines DL1, DL2, ..., DLm, and "BLy" represents the bit lines BL1, BL2, ..., BLn.
[0133] (4-1) Data writing operation
[0134] For example, when charges are injected into the charge storage layer EC of the memory cell MC11, Figure 6 As shown in the "Selection Column" column of "Write", a charge storage gate voltage of 12V is applied from the memory gate line ML to the memory gate MG of the memory cell MC11, and a charge storage gate voltage of 12V is applied to the semiconductor substrate S1 ( Figure 6 , marked by “Back”) a substrate voltage of 0 V is applied.
[0135] In addition, at this time, a gate cutoff voltage of 0V is applied from the second selection gate line SGL to the second selection gate SG of the memory cell MC11, and a source cutoff voltage of 0V is applied from the source line SL to the source region of the memory cell MC11. As a result, in the second selection gate structure 4, a source-side non-conductive region is formed in the fin S2, thereby blocking the electrical connection between the source region and the channel layer forming carrier region (region inducing carrier when forming the channel layer) in the fin S2 surrounded by the memory gate structure 2. Therefore, the second selection gate structure 4 prevents the application of voltage from the source line SL to the channel layer forming carrier region of the memory gate structure 2.
[0136] On the other hand, a first selection gate voltage of 1.5 V is applied to the first selection gate DG from the first selection gate line DL1, and a charge storage bit voltage of 0 V is applied to the drain region of the memory cell MC11 from the bit line BL1. As a result, in the first selection gate structure 3, a drain-side conductive region is formed in the fin S2, thereby electrically connecting the drain region and the channel layer-forming carrier region of the memory gate structure 2.
[0137] In the memory gate structure 2, the carrier region is electrically connected to the drain region through the channel layer formation, so that carriers are induced in the channel layer formation carrier region, and a channel layer of 0V, which is the same as the charge storage bit voltage, is formed on the surface of the fin S2 through the carriers. Therefore, in the memory cell (hereinafter referred to as the write selection memory cell) MC11 for writing data, a large voltage difference (12V) of 12V is generated between the memory gate MG and the channel layer in the memory gate structure 2, and charges can be injected into the charge storage layer EC through the quantum tunnel effect generated thereby, thereby achieving a state of writing data.
[0138] (4-2) Data non-write operation
[0139] For example, when a charge storage gate voltage required for injecting charges into the charge storage layer EC is applied to the memory gate MG of the memory cell MC12, in order to prevent the injection of charges into the charge storage layer EC in the memory cell MC12, a voltage of 1.5V is applied from the first selection gate line DL1 to the first selection gate DG, and a voltage of 1.5V is applied from the bit line BL2 to the drain region, thereby forming a non-conducting drain-side non-conducting region in the region spanned by the first selection gate structure 3 in the fin S2. Thus, in the memory cell MC12 where data is not written (hereinafter referred to as a write non-selected memory cell), the electrical connection between the carrier region and the drain region formed by the channel layer in the fin S2 surrounded by the memory gate structure 2 is blocked by the first selection gate structure 3.
[0140] In addition, at this time, in the write non-selected memory cell MC12, a voltage of 0V is applied from the second selection gate line SGL to the second selection gate SG, and a voltage of 0V is applied from the source line SL to the source region, thereby forming a source-side non-conductive region in a non-conductive state in the region spanned by the second selection gate structure 4 in the fin S2. Thus, in the write non-selected memory cell MC12, the electrical connection between the channel layer forming carrier region and the source region in the fin S2 surrounded by the memory gate structure 2 is blocked by the second selection gate structure 4.
[0141] Therefore, when writing into the non-selected memory cell MC12, a depletion layer is formed in the carrier region of the channel layer within the fin S2 surrounded by the memory gate structure 2. Based on the charge storage gate voltage, the surface potential of the fin S2 surrounded by the memory gate structure 2 rises, and the voltage difference between the memory gate MG and the surface of the fin S2 becomes smaller.
[0142] In particular, in this embodiment, in the memory cell MC, as Figure 4A and Figure 4B As shown, the fin S2 is formed in such a way that the height Hfin of the electrode protrusion in the fin S2 and the width Wfin of the fin S2 satisfy the relationship Hfin>Wfin, so when a depletion layer is formed along the two side surfaces and the upper surface in the fin S2 surrounded by the memory gate structure 2, the depletion layer of a predetermined thickness formed along one side surface in the fin S2 and the depletion layer of a predetermined thickness formed along the other side surface arranged opposite to the one side surface are integrated, and a depletion layer can be formed in the entire fin S2. Therefore, when writing to the non-selected memory cell MC12, the voltage difference that causes the quantum tunneling effect is not generated between the memory gate MG and the fin S2 through the depletion layer, thereby preventing the injection of charges into the charge storage layer EC.
[0143] At this time, in the non-selected memory cell MC12, the potential of the surface of the fin S2 directly below the memory gate structure 2 is prevented from reaching the first selection gate insulating film 17a, 18a of the first selection gate structure 3 by the depletion layer D formed in the fin S2 surrounded by the memory gate structure 2. Figure 3A ), the second selection gate insulating films 17b and 18b of the second selection gate structure 4 ( Figure 3A ).
[0144] Thus, in the first selection gate structure 3, even if the drain region 12a ( Figure 3A ) forms a thin film thickness of the first selection gate insulating film 17a, 18a to match the bit voltage of the low voltage applied, and since the potential on the surface of the fin S2 directly below the memory gate structure 2 is blocked by the depletion layer, the insulation breakdown of the first selection gate insulating film 17a, 18a caused by the potential on the surface of the fin S2 can be prevented.
[0145] Similarly, in the second selection gate structure 4, even if the source SL is connected to the source region 12b ( Figure 3A ) The second selection gate insulating film 17b, 18b is formed to be thin in thickness to match the low voltage source voltage applied. Since the potential on the surface of the fin S2 directly below the memory gate structure 2 is blocked by the depletion layer, the insulation breakdown of the second selection gate insulating film 17b, 18b caused by the potential on the surface of the fin S2 can be prevented.
[0146] (4-3) Data reading operation
[0147] also, Figure 6 In the data reading operation represented by the "Read" column, for example, when the memory cell MC11 is taken as the object of reading, the bit line BL1 connected to the memory cell MC11 is precharged to, for example, 1.5V, and the source line SL is 0V. Thus, when there is charge stored in the charge storage layer EC of the memory cell MC11 for reading data (when data is written), the fin S2 directly under the memory gate structure 2 becomes non-conductive, and the electrical connection between the drain region and the source region can be blocked. Thus, in the memory cell MC11 for reading data, the 1.5V reading voltage of the bit line BL1 connected to the drain region can be kept as it is.
[0148] On the other hand, when no charge is stored in the charge storage layer EC of the memory cell MC11 for reading data (when no data is written), the fin S2 directly below the memory gate structure 2 becomes conductive, and the drain region and the source region are electrically connected, and as a result, the source line LS of 0V and the bit line BL1 of 1.5V are electrically connected through the memory cell MC11. As a result, in the nonvolatile semiconductor memory device 1, the read voltage of the bit line BL1 connected to the memory cell MC11 for reading data is applied to the source line SL of 0V, so that the read voltage of 1.5V applied to the bit line BL1 decreases.
[0149] Therefore, in the nonvolatile semiconductor memory device 1, by detecting whether the read voltage of the bit line BL1 changes, it is possible to perform a read operation of data on whether there is charge stored in the charge storage layer EC of the memory cell MC11. In addition, a non-read voltage of 0V can be applied to the bit line BL2 connected only to the memory cells MC12, MC22, ..., MCm2 that do not read data.
[0150] (4-4) Data erasure action
[0151] Incidentally, the erasing operation ( Figure 6 In “erase”, a memory gate voltage of -12V is applied from the memory gate line ML to the memory gate MG, whereby the charge in the charge storage layer EC is extracted to the fin S2 which becomes 0V through the semiconductor substrate S1 to erase the data.
[0152] (5) Regarding the potential of the high voltage charge storage gate voltage applied to the memory gate to write the non-selected memory cell
[0153] Here, Figure 71 is a schematic diagram showing the potential state of the memory gate structure 2 and the fin S2 when a high voltage charge storage gate voltage Vg is applied to the memory gate MG in a memory cell (write non-selected memory cell) MC where data is not written. Figure 7 The cross-sectional view shown is similar to the Figure 1 and Figure 2 The cross-sectional structure of the C-C' part Figure 4A Similarly, the cross-sectional structure of the fin S2 position where the memory gate structure 2 is provided is shown.
[0154] also, Figure 7 The curve Vx in shows the potential state of a memory cell (comparative example) in which a memory gate structure, a first selection gate structure, and a second selection gate structure are provided on a flat surface of a semiconductor substrate without a fin portion. In the memory cell of the comparative example, when a high-voltage charge storage gate voltage is applied to the memory gate MG, as described above, a depletion layer D can be formed on the semiconductor substrate directly under the memory gate structure by making the semiconductor substrate directly under the first selection gate structure and the semiconductor substrate directly under the second selection gate structure respectively non-conductive, and as a result, the potential difference Vono1 generated between the memory gate and the semiconductor substrate can be reduced to prevent data from being written.
[0155] Even in the memory cell of this comparative example, the voltage in the three-layer structure portion ONO of the upper memory gate insulating film, the charge storage layer, and the lower memory gate insulating film provided in the memory gate structure decreases, and further the voltage value in the depletion layer D also decreases as it moves away from the substrate surface, and can become a substrate voltage of 0 V. However, in the memory cell of the comparative example, since the fin portion S2 is not formed, the voltage of the three-layer structure portion ONO of the upper memory gate insulating film, the charge storage layer, and the lower memory gate insulating film and the potential change based on the depletion layer D are determined by the impurity concentration in the semiconductor substrate directly below the memory gate structure.
[0156] That is, in the memory cell of the comparative example, the more the impurity concentration of the semiconductor substrate directly below the memory gate structure is reduced, the deeper the depletion layer D can be formed when the high voltage charge storage gate voltage Vg is applied to the memory gate MG. Therefore, in the memory cell of the comparative example, the deeper the depletion layer D is, the slower the potential change of the three-layer structure ONO of the upper memory gate insulating film, the charge storage layer, and the lower memory gate insulating film becomes, so that the potential difference between the memory gate and the surface of the semiconductor substrate can be reduced accordingly, and the occurrence of interference can be suppressed.
[0157] However, in the memory cell of the comparative example, when the impurity concentration in the semiconductor substrate directly below the memory gate structure is reduced, a short circuit or leakage may occur between the drain region located on the surface of the semiconductor substrate adjacent to the first selection gate structure and the source region located on the surface of the semiconductor substrate adjacent to the second selection gate structure due to the short channel effect. Therefore, it is necessary to increase the distance between the drain region and the source region, and accordingly, miniaturization based on expansion cannot be achieved.
[0158] As described above, in the memory cell of the comparison example in which the memory gate structure, the first selection gate structure, and the second selection gate structure are arranged on a flat surface of a semiconductor substrate, there is a trade-off between suppressing interference by reducing the impurity concentration in the semiconductor substrate directly below the memory gate structure and achieving miniaturization by bringing the drain region and the source region closer together.
[0159] Here, for example, in a memory cell of a comparative example in which the distance between the drain region and the source region is set at the same distance as the distance between the drain region 12a and the source region 12b of the memory cell MC, as shown in FIG. Figure 7 As shown, when a high-voltage charge storage gate voltage Vg is applied to the memory gate MG, although the charge storage gate voltage Vg decreases in the three-layer structure portion ONO of the upper memory gate insulating film, the charge storage layer, and the lower memory gate insulating film, the distance between the drain region and the source region is a specified value, and the impurity concentration in the semiconductor substrate cannot be reduced to below a predetermined value, so it is difficult to set the potential difference Vono1 generated between the memory gate and the semiconductor substrate to below a predetermined value. Therefore, in the memory cell of the comparative example, when miniaturization is achieved by bringing the drain region and the source region closer, a large potential difference Vono1 is generated between the memory gate and the semiconductor substrate, and the occurrence of interference cannot be suppressed.
[0160] In contrast, in the memory cell MC, when no data is written, Figure 7 As shown, when the charge storage gate voltage Vg (for example, Vg=12V) is applied to the memory gate MG, although the voltage in the three-layer structure part ONO of the upper memory gate insulating film 14, the charge storage layer EC and the lower memory gate insulating film 13 decreases, and a potential difference Vono is generated between the memory gate MG and the upper surface of the fin S2, a depletion layer is formed in the entire fin S2 surrounded by the memory gate structure 2 at this time, and the voltage in the depletion layer also decreases slowly, and the substrate voltage can become 0V near the lower end surface of the fin S2.
[0161] Furthermore, in the memory cell MC where no data is written, the voltage difference Vono of the three-layer structure portion ONO of the upper memory gate insulating film 14, the charge storage layer EC and the lower memory gate insulating film 13 and the potential change based on the depletion layer can be controlled by the height of the fin S2 in the area surrounded by the memory gate structure 2 (the distance from the upper surface of the fin S2 to the lower end surface) and the width of the fin S2 (the distance between the two side surfaces of the fin S2 relatively arranged in the x direction of the memory gate structure 2 across the fin S2).
[0162] Specifically, when a depletion layer is formed along the two side surfaces and the upper surface of the fin S2 surrounded by the memory gate structure 2, the height and width of the fin S2 are set in such a way that a depletion layer of a predetermined thickness formed along one side surface within the fin S2 and a depletion layer of a predetermined thickness formed along another side surface arranged opposite to the one side surface are integrated, and a depletion layer is formed within the entire fin S2.
[0163] Thus, in the memory cell MC, even if the impurity concentration in the fin S2 is increased to bring the drain region 12a and the source region 12b closer to achieve miniaturization, since a depletion layer can be formed in the entire fin S2, the potential difference Vono generated between the memory gate MG and the fin S2 can be reduced. Therefore, in the memory cell MC, miniaturization can be achieved by bringing the drain region 12a and the source region 12b closer, and the potential difference Vono generated between the memory gate MG and the fin S2 is reduced, and the electric field applied to the depletion layer is further reduced, thereby suppressing the occurrence of interference.
[0164] (6) Method for manufacturing nonvolatile semiconductor memory device
[0165] The nonvolatile semiconductor memory device 1 having the above structure is manufactured by the following manufacturing process. Figure 8 As shown, by processing a fin forming substrate (not shown) composed of Si, for example, using a patterned hard mask 32a, fins S2 are formed at predetermined intervals on the surface of a plate-shaped semiconductor substrate S1. Next, an insulating layer composed of an insulating material is formed in a manner covering the hard mask 32a, the semiconductor substrate S1, and the fins S2, and then the insulating layer is flattened by a flattening process such as chemical mechanical polishing (CMP), forming an insulating layer Isa whose surface is consistent with the hard mask 32a on the upper surface of the fin S2.
[0166] Next, by processing the insulating layer ISa, as Fig. 9A As shown, the fin S2 is formed to protrude from the surface by a predetermined height of the insulating layer IS, and then, as shown in FIG. Fig. 9A and show Fig. 9AThe cross-sectional structure of the E-E' section Fig. 9B As shown, a layered lower memory gate insulating film forming layer 13a, a layered charge storage layer forming layer ECa and a layered first dummy electrode layer 34 are formed, and the surface of the first dummy electrode layer 34 is flattened by a flattening process such as CMP.
[0167] Next, by processing the first dummy electrode layer 34, the charge storage layer forming layer ECa and the lower memory gate insulating film forming layer 13a, as shown in FIG. Fig. 9C and show Fig. 9C The cross-sectional structure of the E-E' section Fig.9D As shown, a virtual memory gate DM, a charge storage layer EC and a lower memory gate insulating film 13 are formed to extend across the fin S2, and then a layered insulating film is formed and etched back to form sidewall-like sidewall spacers 6 and 7 along the side walls of the stacked virtual memory gate DM, the charge storage layer EC and the lower memory gate insulating film 13 (sidewall spacer formation process).
[0168] Next, a layered selection gate insulating film and a layered second virtual electrode layer are sequentially formed in a manner of covering the insulating layer IS, the fin S2, the virtual memory gate DM and the sidewall spacers 6 and 7 exposed to the outside, and then etched back as in the case of Fig.9D The corresponding parts are marked with the same reference numerals. Fig. 10A As shown, a virtual first selection gate DD and a virtual second selection gate DS are formed in the shape of sidewalls along the sidewall spacers 6 and 7. Next, the selection gate insulating film is removed except for the area covered by the virtual first selection gate DD and the virtual second selection gate DS, and the selection gate insulating film is left at the lower parts of the virtual first selection gate DD and the virtual second selection gate DS, respectively, to form a first selection gate insulating film 17a and a second selection gate insulating film 17b.
[0169] Next, a layered insulating film is formed to cover the insulating layer IS, the fin S2, the dummy memory gate DM, the dummy first selection gate DD, and the dummy second selection gate DS exposed to the outside, and then etched back, as in the Fig. 10A The corresponding parts are marked with the same reference numerals. Fig. 10B As shown, a sidewall portion 22a is formed along the sidewalls of the virtual first selection gate DD and the first selection gate insulating film 17a on one side, and a sidewall portion 22b is formed along the sidewalls of the virtual second selection gate DS and the second selection gate insulating film 17b on the other side (virtual selection gate formation process).
[0170] Next, a drain region 12a and a source region 12b with a predetermined film thickness made of SiGe or the like are formed on the surface of the fin S2 adjacent to the sidewalls 22a and 22b, for example, by epitaxial growth. Next, a layered interlayer insulating layer is formed to cover the insulating layer IS, the fin S2, the dummy memory gate DM, the dummy first selection gate DD, the dummy second selection gate DS, the drain region 12a, and the source region 12b exposed to the outside, and then the surface of the interlayer insulating layer is polished and flattened by a flattening process such as CMP.
[0171] Therefore, if Fig. 10B The corresponding parts are marked with the same reference numerals. Fig. 10C As shown, an interlayer insulating layer 25 is formed whose surface is planarized and each surface of the dummy memory gate DM, the dummy first selection gate DD, and the dummy second selection gate DS is exposed from the surface to the outside (dummy electrode formation step). Next, the dummy memory gate DM, the dummy first selection gate DD, and the dummy second selection gate DS exposed from the surface of the interlayer insulating layer 25 are removed by dry etching.
[0172] Therefore, if Fig. 9C The corresponding parts are marked with the same reference numerals. Fig.11A As shown, a virtual electrode removal space ER1 is formed in the region where the virtual memory gate DM is formed. In addition, as shown Fig.11A The cross-sectional structure of the E-E' section Fig. 11B As shown, dummy electrode removal spaces ER2 and ER3 are also formed in the region where the dummy first selection gate DD and the dummy second selection gate DS are formed (dummy electrode removal step).
[0173] Next, a layered insulating film made of an insulating material such as a high-k material is formed in the dummy electrode removal spaces ER1, ER2, and ER3. Figure 3A As shown, in the virtual electrode removal space ER1, a layered upper memory gate insulating film 14 is formed on the charge storage layer EC, a wall-shaped memory gate side wall insulating film 15a is formed along one side wall partition 6, and a wall-shaped memory gate side wall insulating film 15b is formed along the other side wall partition 7.
[0174] In addition, in the virtual electrode removal space ER2, an upper first selection gate insulating film 18a is formed on the first selection gate insulating film 17a located on the fin portion S2, a wall-shaped side wall insulating film 20a is formed along the side wall portion 22a, and a wall-shaped first selection gate side wall insulating film 19a is formed along the side wall partition 6.
[0175] Furthermore, in the virtual electrode removal space ER3, an upper second selection gate insulating film 18b is formed on the second selection gate insulating film 17b located on the fin S2, a wall-shaped side wall insulating film 20b is formed along the side wall portion 22b, and a wall-shaped second selection gate side wall insulating film 19b is formed along the other side wall partition 7.
[0176] Next, a metal electrode layer made of a metal material such as aluminum (Al), titanium aluminum (TiAl), tantalum carbide (TaC), or tantalum silicon nitride (TaSiN) is formed on the surface of the interlayer insulating layer 25. Thus, the virtual memory gate DM, the virtual first selection gate DD, and the virtual second selection gate DS are removed, and the metal electrode layer is buried in the virtual electrode removal spaces ER1, ER2, and ER3 formed with the layered insulating film, respectively, and then the surface of the metal electrode layer is polished by a planarization process such as CMP, and the surface of the metal electrode layer is planarized to match the surface of the interlayer insulating layer 25.
[0177] Therefore, if Figure 3A As shown, a metal electrode layer is buried in the space where the virtual memory gate DM is formed to form a memory gate MG, a metal electrode layer is buried in the space where the virtual first selection gate DD is formed to form a first selection gate DG, and a metal electrode layer is buried in the space where the virtual second selection gate DS is formed to form a second selection gate SG, thereby forming a memory cell MC including a memory gate structure 2, a first selection gate structure 3, and a second selection gate structure 4 (metal gate formation process). After that, in addition to the process of forming the upper interlayer insulating layer, a metal gate is formed at a predetermined position of the interlayer insulating layer 25 and the upper interlayer insulating layer. Figure 2 The nonvolatile semiconductor memory device 1 is manufactured by the steps of forming various connectors such as the first select gate connector DGC, the second select gate connector SGC, and the memory gate connector MCG.
[0178] (7) Function and effect
[0179] According to the above structure, in the memory cell MC, the fin S2 is provided on the semiconductor substrate S1 in a manner protruding from the insulating layer IS, and the memory gate structure 2, the first selection gate structure 3, and the second selection gate structure 4 are formed on the insulating layer IS in a manner spanning the fin S2. In addition, in the memory cell MC, the drain region 12a connected to the bit line BL is provided on the surface of the fin S2 adjacent to the first selection gate structure 3 in a manner insulated from the first selection gate DG, and on the other hand, the source region 12b connected to the source line SL is provided on the surface of the fin S2 adjacent to the second selection gate structure 4 in a manner insulated from the second selection gate SG, and the first selection gate structure 3, the memory gate structure 2, and the second selection gate structure 4 are provided between the drain region 12a and the source region 12b.
[0180] In a memory cell MC having such a structure, when preventing data writing by blocking the injection of charges into the charge storage layer EC, the electrical connection between the fin S2 surrounded by the memory gate structure 2 and the bit line BL is blocked by the first selection gate structure 3, and the electrical connection between the fin S2 surrounded by the memory gate structure 2 and the source line SL is blocked by the second selection gate structure 4, so that the depletion layer formed along the surface of the fin S2 surrounded by the memory gate structure 2 is integrated to form a depletion layer in the entire fin S, and the depletion layer can reduce the potential difference between the memory gate DG and the fin S2.
[0181] Therefore, in the memory cell MC, even if the impurity concentration in the fin S2 is increased, the drain region 12a and the source region 12b are brought close to each other on the surface of the fin S2 to achieve miniaturization, since the shape (height and width) of the fin S2 is set in such a way that a depletion layer is formed in the entire fin S2, the potential difference between the memory gate DG and the fin S2 becomes smaller, and the electric field of the depletion layer is further reduced, thereby suppressing the occurrence of interference. Therefore, the memory cell MC can achieve miniaturization while suppressing the occurrence of interference.
[0182] In addition, each gate width of the memory gate structure 2, the first selection gate structure 3 and the second selection gate structure 4 is formed in a "U" shape along the two side surfaces and the surface of the fin S2, so that a part of the gate width can be replaced by the height of the fin S2. Therefore, although the height is increased to the height of the fin S2, the formation area of each gate width can be narrowed for the x direction across the fin S2, thereby achieving miniaturization.
[0183] Furthermore, in the memory cell MC, when no data is written, the voltage values of the bit line BL and the source line SL can be reduced to a voltage value that causes only the fins S2 directly below the first selection gate structure 3 and the second selection gate structure 4 to be in a non-conducting state without being restricted by the high voltage charge storage gate voltage applied to the memory gate MG. Therefore, in matching with the voltage reduction of the bit line and the source line, the thickness of the first selection gate insulating film 17a, 18a of the first selection gate structure 3 and the second selection gate insulating film 17b, 18b of the second selection gate structure 4 can be thinned, thereby achieving miniaturization accordingly.
[0184] Furthermore, in the nonvolatile semiconductor memory device 1 in which a plurality of memory cells MC sharing a memory gate line ML are arranged in a matrix, the potential difference Vono between the memory gate MG and the fin S2 generated in the memory cell MC in which no data is written is small, and the electric field of the depletion layer is also small, so even if a high-voltage charge storage gate voltage is applied to the memory gate MG multiple times via the memory gate line ML, interference in the memory cell MC in which no data is written can be suppressed. Therefore, in the nonvolatile semiconductor memory device 1, for example, even if the memory cells MC arranged in 128 or more rows and 1024 or more columns share a memory gate line ML, interference can be suppressed, thereby significantly increasing the number of memory cells that can be processed in batches.
[0185] (8) Other Implementation Methods
[0186] (8-1) Structure of a memory cell when the memory gate, the first selection gate, and the second selection gate are formed of a conductive material other than a metal material
[0187] Furthermore, in the above-mentioned embodiment, the memory gate MG, the first selection gate DG and the second selection gate SG are described as being formed of metal materials, but the present invention is not limited thereto, and the memory gate, the first selection gate and the second selection gate may also be formed of various other conductive materials such as polysilicon.
[0188] Here, the following describes an embodiment in which the memory gate MG, the first selection gate DG, and the second selection gate SG are formed of a conductive material other than a metal material (for example, polysilicon). In this case, the plane layout of the nonvolatile semiconductor memory device is different from Figure 2 The structure shown is the same, but Figure 2 The cross-sectional structure of the A-A' portion becomes as follows Fig. 12A The structure shown, Figure 2 The cross-sectional structure of the B-B' portion becomes as follows Fig. 12B The structure shown.
[0189] As Figure 3AThe corresponding parts are marked with the same reference numerals. Fig. 12A As shown, in the memory cell MC1, on the surface of the fin S2, an impurity diffusion region formed by the injection of impurities is arranged at a predetermined interval as a drain region 40a and a source region 40b, and the drain region 40a is connected to a bit line (not shown), and the source region 40b is connected to a source line (not shown).
[0190] In addition, the memory cell MC1 has a memory gate structure 2a in which a lower memory gate insulating film 13, a charge storage layer EC, an upper memory gate insulating film 14a and a memory gate MG1 are stacked in sequence on the fin S2, the lower memory gate insulating film 13 and the upper memory gate insulating film 14a are formed of the same insulating material (for example, silicon oxide (SiO, SiO2), etc.), and the memory gate MG1 is formed of polycrystalline silicon.
[0191] Between the drain region 40a and the memory gate structure 2a, a first selection gate structure 3a having a structure in which a first selection gate DG1 made of polysilicon is stacked on a first selection gate insulating film 17a is provided. The first selection gate structure 3a is adjacent to a sidewall spacer 6 formed on one sidewall of the memory gate structure 2a, and is formed in a sidewall shape such that the top of the first selection gate DG1 decreases toward the fin S2 as it moves away from the memory gate MG1.
[0192] Furthermore, between the source region 40b and the memory gate structure 2a, a second selection gate structure 4a having a structure in which a second selection gate SG1 made of polysilicon is stacked on a second selection gate insulating film 17b is provided. The second selection gate structure 4a is formed adjacent to a sidewall spacer 7 formed on the other sidewall of the memory gate structure 2a, and is formed in a sidewall shape in which the top of the second selection gate SG1 decreases toward the fin S2 as it moves away from the memory gate MG1.
[0193] Such a memory gate structure 2a, a first selection gate structure 3a, and a second selection gate structure 4a are also formed in a manner spanning the fin S2, even on the insulating layer IS where the fin S2 is not formed, as in Figure 3B The corresponding parts are marked with the same reference numerals. Fig. 12B As shown, you can set Fig. 12A The memory gate structure 2a, the first select gate structure 3a and the second select gate structure 4a are shown.
[0194] Furthermore, in such a memory cell MC1, in the manufacturing process, the first selection gate DG1 and the second selection gate SG1 in the form of sidewalls are formed along the sidewalls of the memory gate structure 2a by, for example, etching back the layered electrode forming layer. Therefore, the memory gate structure 2a, the first selection gate structure 3a, the second selection gate structure 4a, and the sidewall spacers 6 and 7 can be formed with the same thickness at the position where the fin S2 is formed and the position where the fin S2 is not formed. Therefore, at the position where the fin S2 is formed, the memory gate structure 2a, the first selection gate structure 3a, the second selection gate structure 4a, and the sidewall spacers 6 and 7 can protrude by the height of the fin S2 compared to the position where the fin S2 is not formed.
[0195] Incidentally, in the memory cell MC1, as in the above-described embodiment, the distance Hfin from the upper surface of the fin S2 to the bottom surface of the memory gate MG1 on the insulating layer IS (the in-electrode protrusion height of the fin S2) and the width Wfin of the fin S2 in the x direction of the memory gate structure 2a, the first selection gate structure 3a, and the second selection gate structure 4a are equal. Figure 4A and Figure 4B ) is formed in a way that satisfies the relationship Hfin﹥Wfin.
[0196] In addition, in the memory cell MC1, the gate length L1 of the first selection gate DG1 extending in the y direction of the fin S2, the gate length L2 of the second selection gate SG1, and the width Wfin of the fin S2 are formed in a manner to satisfy the relationship of L1≦1.5·Wfin and L2≦1.5·Wfin. Furthermore, even such a memory cell MC1 can implement data writing, data non-writing, data reading, and data erasing according to the above-mentioned "(4) Voltages during various operations of the non-volatile semiconductor memory device".
[0197] Even with the memory cell MC1 having the above-described structure, the same effects as those of the above-described embodiment can be obtained, and the occurrence of disturbance can be suppressed while achieving miniaturization.
[0198] (8-2) Structure of a memory cell having a cap insulating film provided on the upper surface of the fin
[0199] In the above-mentioned embodiment, the memory cell MC in which the lower memory gate insulating film 13 is provided on the upper surface of the fin S2 is described, but the present invention is not limited to this. The memory cell may also be a memory cell in which a cover insulating film having a predetermined thickness is provided on the upper surface of the fin S2, and a lower memory gate insulating film 13 is provided on the cover insulating film.
[0200] in this case, Fig.13A nonvolatile semiconductor memory device 31 according to another embodiment is shown. In the nonvolatile semiconductor memory device 31, a memory cell MC2 including a memory gate structure 32 having a cap insulating film (not shown) is provided. Figure 2 The memory cell MC shown has the same external structure, but a cap insulating film is provided on the upper surface of the fin portion S2 covered by the memory gate structure 32 .
[0201] Here, Fig.14A Show Fig.13 The cross-sectional structure of the A-A' section, Fig. 14B Show Fig.13 The cross-sectional structure of the B-B' part, Fig.15A Show Fig.13 The cross-sectional structure of the C-C' part, Fig. 15B Show Fig.13 The cross-sectional structure of the D-D' portion. Fig.14A and Fig.15A As shown, the memory gate structure 32 has a cover insulating film 35 with a predetermined film thickness arranged on the upper surface of the fin S2, and a structure in which a lower memory gate insulating film 13, a charge storage layer EC, an upper memory gate insulating film 14 and a memory gate MG are stacked in sequence on the cover insulating film 35.
[0202] By the way, in this embodiment, a memory gate structure 32 in which a lower memory gate insulating film 13 is formed on a cover insulating film 35 is described, but the present invention is not limited to this. The memory gate structure may be, for example, a cover insulating film 35 provided as a lower memory gate insulating film 13, and a charge storage layer EC, an upper memory gate insulating film 14 and a memory gate MG are sequentially stacked on the cover insulating film 35, so that the cover insulating film 35 plays the role of a lower memory gate insulating film 35.
[0203] On the other hand, Fig. 14B As shown, the memory gate structure 32 has the following structure: the cap insulating film 35 is not formed at the position where the fin S2 is not formed, the lower memory gate insulating film 13 is formed on the insulating layer IS, and the charge storage layer EC, the upper memory gate insulating film 14 and the memory gate MG are sequentially stacked on the lower memory gate insulating film 13. As described above, the memory gate structure 32 has a structure in which the cap insulating film 35 is provided only in the region of the upper surface where the fin S2 is arranged, and the upper surface of the fin S2 is covered by the cap insulating film 35.
[0204] The cover insulating film 35 is made of insulating material such as silicon oxide (SiO, SiO2, SiN), and has a thickness of, for example, 4 nm or more. The cover insulating film 35 may have a structure in which different types of insulating materials are stacked in addition to a structure in which one type of insulating material is layered.
[0205] And, if Fig.14A , Fig. 14B and Fig. 15B As shown, the first selection gate structure 3 has the following structure: it does not have a cap insulating film 35, and the first selection gate insulating films 17a, 18a and the first selection gate DG are sequentially stacked on the fin S2 or the insulating layer IS. In addition, similarly, the second selection gate structure 4 also has the following structure: it does not have a cap insulating film 35, and the second selection gate insulating films 17b, 18b and the second selection gate SG are sequentially stacked on the fin S2 or the insulating layer IS.
[0206] Incidentally, as a method for manufacturing the memory gate structure 32 including such a cap insulating film 35, for example, in a fin formation step, Figure 8 As shown, the hard mask 32a used when processing the plate-shaped fin forming substrate (not shown) to form the fin S2 is formed of an insulating material, and the hard mask 32a is left as it is as a cover insulating film 35. In the next first dummy electrode forming step, a layered lower memory gate insulating film forming layer 13a, a charge storage layer forming layer ECa and a first dummy electrode layer 34 are sequentially stacked on the insulating layer IS and the cover insulating film 35 covering the upper surface of the fin S2 (see FIG. Fig. 9A and Fig. 9B ).
[0207] Furthermore, at this time, when the lower memory gate insulating film forming layer 13a is formed by an oxidation method, there may be a case where the lower memory gate insulating film forming layer 13a is not formed on the cover insulating film 35. In this case, the finally formed memory gate structure may have the following structure: the cover insulating film 35 is provided as the lower memory gate insulating film 13, and the charge storage layer EC, the upper memory gate insulating film 14 and the memory gate MG are sequentially stacked on the cover insulating film 35.
[0208] Furthermore, in the next virtual memory gate structure formation process, the lower memory gate insulating film forming layer 13a, the charge storage layer forming layer ECa and the first virtual electrode layer 34 are patterned to form a virtual memory gate structure 32 on the insulating layer IS in a manner that spans the fin S2 on the upper surface covered by the cover insulating film 35. Then, the memory cell MC2 can be manufactured through the above-mentioned sidewall spacer formation process, virtual selection gate formation process, virtual electrode exposure process, metal gate formation process, etc.
[0209] In the above structure, in addition to obtaining the same effects as those of the above-mentioned embodiment, the memory cell MC2 has a cover insulating film 35 covering the upper surface of the fin S2 provided on the memory gate structure 32. Thus, for example, during data writing, the cover insulating film 35 can prevent the electric field from being concentrated around the sharp corners where the upper surface of the fin S2 intersects the side surface. Thus, it is possible to prevent poor writing caused by the electric field only being concentrated around the corners and unable to inject charges into the charge storage layer film along the sidewalls of the fin during writing.
[0210] Incidentally, in the above-mentioned embodiment, the memory gate structure 32 in the memory cell MC2 in which the memory gate MG, the first selection gate DG, and the second selection gate SG are formed of a metal material is described as being provided with a cap insulating film 35 covering the upper surface of the fin S2, but the present invention is not limited thereto. Fig. 12A and Fig. 12B As shown, a cap insulating film 35 covering the upper surface of the fin S2 may be provided in the memory gate structure 2a in the memory cell MC1 in which the memory gate MG1, the first selection gate DG1, and the second selection gate SG1 are formed of a conductive material such as polysilicon. Even in this case, in the memory cell MC1, during the data erasing operation, the cap insulating film 35 can prevent the electric field from being concentrated around the sharp corners where the upper surface of the fin S2 intersects with the side surface.
[0211] (8-3) Others
[0212] Furthermore, the present invention is not limited to the above-mentioned embodiments, and various modifications can be made within the scope of the gist of the present invention. For each embodiment, for example, various voltage values other than the voltage value of the above-mentioned "(4) Voltages during various operations of the non-volatile semiconductor storage device" can be applied.
[0213] in addition, Figure 3A and Figure 3B In the memory cell MC shown, the case where the first selection gate insulating film 17a, 18a and the second selection gate insulating film 17b, 18b having a two-layer structure are used as the first selection gate insulating film and the second selection gate insulating film is described, but the present invention is not limited to this. The first selection gate insulating film 17a, 18a may be constructed as a one-layer structure on either side, or the second selection gate insulating film 17b, 18b may be constructed as a one-layer structure on either side.
[0214] Furthermore, in the above-mentioned embodiment, the memory gate MG ( Figure 3A and Figure 3B) is described, but the present invention is not limited to this. It can also be a memory gate set along any side wall of any side of the memory gate sidewall insulating film 15a, 15b, or it can be a memory gate without the memory gate sidewall insulating film 15a, 15b.
[0215] Furthermore, in the above-mentioned embodiment, the first selection gate sidewall insulating film 19a and the sidewall insulating film 20a are provided along the sidewall of the first selection gate DG ( Figure 3A and Figure 3B ) is described, but the present invention is not limited to this. It can also be a first selection gate in which any one of the first selection gate sidewall insulating film 19a or the sidewall insulating film 20a is set along any side wall, or it can be a first selection gate in which neither the first selection gate sidewall insulating film 19a nor the sidewall insulating film 20a is set.
[0216] Furthermore, in the above-mentioned embodiment, the second selection gate sidewall insulating film 19b and the sidewall insulating film 20b are provided along the sidewalls of the second selection gate SG ( Figure 3A and Figure 3B ) has been described, but the present invention is not limited to this. It can also be a second selection gate in which either the second selection gate sidewall insulating film 19b or the sidewall insulating film 20b is set along any side wall, or it can be a second selection gate in which neither the second selection gate sidewall insulating film 19b nor the sidewall insulating film 20b is set.
[0217] Furthermore, in the above-mentioned embodiment, the upper memory gate insulating film 14, the memory gate sidewall insulating films 15a, 15b, the first selection gate sidewall insulating film 19a, the sidewall sidewall insulating film 20a, the second selection gate sidewall insulating film 19a and the sidewall sidewall insulating film 20b are formed of insulating materials different from the sidewall spacers 6 and 7, but the present invention is not limited thereto, and the upper memory gate insulating film 14, the memory gate sidewall insulating films 15a, 15b, the first selection gate sidewall insulating film 19a, the sidewall sidewall insulating film 20a, the second selection gate sidewall insulating film 19a and the sidewall sidewall insulating film 20b may be formed of the same insulating material as the sidewall spacers 6 and 7. Furthermore, the memory cell may also be formed on the semiconductor substrate S1 and the fin S2 into which N-type impurities are injected.
[0218] Furthermore, in the above-mentioned embodiment, the case where the memory gate MG, the first selection gate DG and the second selection gate SG are formed by using a metal layer is described, but the present invention is not limited to this. It can also be, for example, that a plurality of metal layers composed of different types of metal materials are stacked in sequence to form a memory gate MG, a first selection gate DG and a second selection gate SG having a stacked structure.
[0219] Furthermore, in the above-mentioned embodiment, a case is described in which a drain region 12a and a source region 12b of a predetermined thickness composed of SiGe or the like are formed by, for example, an epitaxial growth method as the drain region and the source region, but the present invention is not limited to this. Impurities may be injected into the surface of the fin S2 to form an impurity diffusion region, and it may be used as the drain region and the source region.
[0220] Furthermore, in the above-mentioned embodiment, the following virtual selection gate forming process is described as a virtual selection gate forming process in the manufacturing method, that is, a selection gate insulating film and a second virtual electrode layer are formed in a manner covering the virtual memory gate structure, and then a virtual first selection gate in a sidewall shape is formed in a manner across the fin portion along a side wall partition located on the side wall of the virtual memory gate structure by etching back, and a virtual second selection gate in a sidewall shape is formed in a manner across the fin portion along another side wall partition located on the side wall of the virtual memory gate structure, and then the selection gate insulating film is processed, and a first selection gate insulating film is provided under the virtual first selection gate, and a second selection gate insulating film is provided under the virtual second selection gate. However, the present invention is not limited to this, and it may also be a virtual selection gate forming process in which a virtual first selection gate and a virtual second selection gate are formed without forming a selection gate insulating film.
[0221] (9) Memory cell in which the film thickness of the insulating layer in the region where the first selection gate structure and the second selection gate structure are provided is reduced
[0222] (9-1) Structure of memory cell
[0223] As Figure 1 The corresponding parts are marked with the same reference numerals. Fig.16 As shown, the difference from the above-mentioned embodiment is that, in the memory cell MC3 provided in the nonvolatile semiconductor memory device 41, the film thickness of the insulating layer IS1 in the region where the first selection gate structure 3 and the second selection gate structure 4 are provided is thinner than the film thickness of the insulating layer IS1 in the region where the memory gate structure 2 is provided. Therefore, here, the insulating layer IS1 is mainly described, and the other structures are the same as those in the above-mentioned embodiment, so the description thereof is omitted.
[0224] The insulating layer IS1 is provided on the surface of the semiconductor substrate S1, and the fin S2 protrudes from the surface of the insulating layer IS1. Since the film thickness of the insulating layer IS1 in the region where the first selection gate structure 3 and the second selection gate structure 4 are provided is thinner than the film thickness of the insulating layer IS1 in the region where the memory gate structure 2 is provided, the insulating layer IS1 is configured to include a layered insulating layer 42 formed on the surface of the semiconductor substrate S1 and a convex insulating layer 43 provided in a manner protruding from the layered insulating layer 42 in the region where the memory gate structure 2 is provided. The convex insulating layer 43 is made of the same insulating material as the layered insulating layer 42 and is formed integrally with the layered insulating layer 42. The convex insulating layer 43 is provided along the memory gate structure 2, and, like the memory gate structure 2, is extended in the x direction whose length direction intersects with the length direction (y direction) of the fin S2.
[0225] Here, with Figure 3A The corresponding parts are marked with the same reference numerals. Fig.17A It is shown Fig.16 A cross-sectional view of the cross-sectional structure of the A-A' portion. Fig.17A As shown, in the region where the fin S2 is provided, the layered insulating layer 42 and the convex insulating layer 43 are not provided ( Fig.16 ), a first selection gate structure 3, a memory gate structure 2 and a second selection gate structure 4 are provided on the surface of the fin S2.
[0226] and Figure 3B The corresponding parts are marked with the same reference numerals. Fig. 17B It is shown Fig.16 The cross-sectional view of the cross-sectional structure of the B-B' portion. Fig. 17B As shown, the first selection gate structure 3 and the second selection gate structure 4 are provided on the layered insulating layer 42 , and the memory gate structure 2 is provided on the convex insulating layer 43 .
[0227] In the region where the fin S2 is not provided ( Fig. 17B ), in the region where the first selection gate structure 3 and the second selection gate structure 4 are provided, the distance from the surface of the semiconductor substrate S1 to the upper surface of the layered insulating layer 42 is H5. On the other hand, in the region where the memory gate structure 2 is provided, the distance from the surface of the semiconductor substrate S1 to the upper surface of the convex insulating layer 43 is H4. The distance H4 is greater than the distance H5, which is the height distance H3 of the convex insulating layer 43.
[0228] Furthermore, the distance H1 from the semiconductor substrate S1 to the lower surface of the memory gate MG is greater than the distance H2 from the semiconductor substrate S1 to each lower surface of the first selection gate DG and the second selection gate SG, which is greater than the distance H3 of the film thickness of the convex insulating layer 43 and, for example, the film thickness of the charge storage layer EC. Therefore, each lower surface position of the first selection gate DG and the second selection gate SG is configured at a position closer to the semiconductor substrate S1 than the lower surface position of the memory gate MG.
[0229] Thus, in the memory cell MC3, the conduction and cutoff of the first selection transistor T1 and the second selection transistor T2 in the side area of the fin S2 can be fully controlled by the first selection gate DG and the second selection gate SG formed at a lower position than the lower surface position of the memory gate MG. That is, when the gate cutoff voltage is applied to the first selection gate DG and the second selection gate SG, the side area of the fin S2 can be reliably covered by the first selection gate DG and the second selection gate SG to become a non-conductive state, so even if the information written to the memory transistor MT is a low threshold voltage Vth, it is possible to prevent leakage current from flowing between the drain area 12a and the source area 12b along the side area of the fin S2. Therefore, in the memory cell MC3, malfunctions during data writing and reading operations can be prevented.
[0230] In addition, this malfunction is described in detail in the above-mentioned "(2) Detailed structure of memory cell" using "a comparative example (not shown) in which the distance H1 is less than the distance H2, and the lower surface positions of the first selection gate DG and the second selection gate SG are arranged at a position higher than the lower surface position of the memory gate MG". Here, the comparative example in which the on and off of the first selection transistor T1 and the second selection transistor T2 cannot be sufficiently controlled is described in the above-mentioned "(2) Detailed structure of memory cell", and therefore its description is omitted.
[0231] Here, the distance H3 as the thickness of the protruding insulating layer 43 is preferably equal to the width Wfin ( Fig.18A and Fig.18B In the fin S2, from the height position of the lower surface of the first selection gate DG and the second selection gate SG to the height of about half (1 / 2) of the width Wfin of the fin S2 toward the upper surface of the fin S2, there may remain an area where the control of the fin S2 (channel layer) cannot be fully performed based on the gate cut-off voltage applied to the first selection gate DG and the second selection gate SG.
[0232] Although the control of the fin S2 based on the gate cutoff voltage can be facilitated by optimizing the impurity concentration in the fin S2, in this case, it is necessary to make the impurity concentration of the fin S2 in the region where the first selection transistor T1 and the second selection transistor T2 are formed different from the optimized value of the impurity concentration of the fin S2 in the region where the memory transistor MT is formed. Here, when the distance H3 as the film thickness of the convex insulating layer 43 is made to be more than 1 / 2 of the width Wfin of the fin S2, the region where the control of the fin S2 (channel layer) based on the gate cutoff voltage cannot be fully performed will be a position lower than the lower surface position of the memory gate MG, so when the information written to the memory transistor MT is a low threshold voltage Vth, a leakage current flows between the drain region 12a and the source region 12b along the side region of the fin S2, thereby preventing the occurrence of malfunction. Therefore, it is not necessary to optimize the impurity concentration of the fin S2 in the region where the first selection transistor T1 and the second selection transistor T2 are formed for the control of the fin S2 based on the gate cutoff voltage, and the occurrence of malfunction can be prevented.
[0233] and Figure 4A The corresponding parts are marked with the same reference numerals. Fig.18A It is shown Fig.16 The memory gate structure 2 is formed on the protruding insulating layer 43 in a manner spanning the fin S2. Thus, in the memory gate structure 2, a channel layer of the memory transistor MT is formed along the surface of the fin S2 surrounded by the memory gate structure 2.
[0234] As described above, in this embodiment, in the memory gate structure 2, the gate width of the memory transistor MT is also the distance along the surface of the fin S2 surrounded by the memory gate structure 2, so a part of the gate width is replaced by the height along the side of the fin S2, so that the formation area of the memory transistor MT can be narrowed accordingly.
[0235] and Figure 4B The corresponding parts are marked with the same reference numerals. Fig.18B It is shown Fig.16 The cross-sectional view of the cross-sectional structure of the D-D' portion. Fig.18B As shown, in the region where the first selection gate structure 3 is formed, the convex insulating layer 43 is not provided on the layered insulating layer 42, and the first selection gate structure 3 is formed on the layered insulating layer 42 in a manner spanning the fin S2. Thus, the first selection gate structure 3 surrounds the surface of the fin S2, and the channel layer of the first selection transistor T1 can be formed along the surface of the fin S surrounded by the first selection gate structure 3.
[0236] As described above, in this embodiment, the gate width of the first selection transistor T1 is also the distance along the surface of the fin S2 surrounded by the first selection gate structure 3, so a part of the gate width is replaced by the height along the side of the fin S2, so that the formation area of the first selection transistor T1 can be narrowed accordingly. Furthermore, the gate width of the first selection transistor T1 is increased to twice the distance H3 of the film thickness of the convex insulating layer 43, so that the on-current of the first selection transistor T1 can be increased accordingly, thereby facilitating high-speed operation. In addition, the second selection structure 4 is also provided on the layered insulating layer 42 in a manner spanning the fin S2, and has the same Fig.18B The first selection gate structure 3 shown has the same structure, so the description thereof is omitted here.
[0237] The memory cell MC3 according to this embodiment is the same as the above-mentioned "(3) Circuit structure of nonvolatile semiconductor memory device" and "(4) Voltages during various operations of nonvolatile semiconductor memory device", and therefore the description thereof is omitted.
[0238] In a method for manufacturing an insulating layer IS1 in which such a protruding insulating layer 43 is arranged on a layered insulating layer 42, for example, in a side wall spacer formation process, when a layered insulating film is formed and then etched back to form side wall spacers 6 and 7, the insulating layer IS exposed to the outside is eliminated by over-etching by a distance H3 corresponding to the height of the protruding insulating layer 43.
[0239] (9-2) Function and Effect
[0240] Through the above structure, in the memory cell MC3 of this embodiment, as in the above embodiment, when preventing the writing of data by preventing the injection of charges into the charge storage layer EC, a depletion layer can be formed in the entire fin S2 surrounded by the memory gate structure 2, and the potential difference between the memory gate DG and the fin S2 can be reduced through the depletion layer. Therefore, in the memory cell MC3, even if the impurity concentration in the fin S2 is increased, the drain region 12a and the source region 12b are brought close to each other on the surface of the fin S2 to achieve miniaturization. Since the shape (height and width) of the fin S2 is set in such a way that a depletion layer is formed in the entire fin S2, the potential difference between the memory gate DG and the fin S2 is reduced, and the electric field of the depletion layer is further reduced, thereby suppressing the occurrence of interference. Therefore, the memory cell MC3 can also be miniaturized while suppressing the occurrence of interference.
[0241] Furthermore, in the memory cell MC3, the insulating layer IS1 is set so that the distance H4 from the semiconductor substrate S1 to the lower surface of the lower memory gate insulating film 13 is greater than the distance H5 from the semiconductor substrate S1 to the lower surfaces of the first selection gate insulating film 17a and the second selection gate insulating film 17b. Thus, in the region where the insulating layer IS1 is set, the lower surface positions of the first selection gate DG and the second selection gate SG can be arranged at positions closer to the semiconductor substrate S1 than the lower surface position of the memory gate MG.
[0242] Therefore, in the memory cell MC3, the first selection gate DG and the second selection gate SG, which are arranged at a lower position than the lower surface position of the memory gate MG, can more reliably control the on and off of the first selection transistor T1 and the second selection transistor T2 in the side area of the fin S2. Therefore, in the memory cell MC3, due to the provision of the convex insulating layer 43, malfunctions during data writing and data reading can be more reliably prevented compared with the above-mentioned embodiment.
[0243] (10) Memory cell having fins on an insulating layer
[0244] (10-1) Structure of memory cell
[0245] As Figure 1 The corresponding parts are marked with the same reference numerals. Fig.19 As shown, in the memory cell MC4 provided in the nonvolatile semiconductor memory device 51, the structure of the fin S3 is different from that in the above-mentioned embodiment. Therefore, the following mainly describes the fin S3, and the other structures are the same as those in the above-mentioned embodiment, so their description is omitted.
[0246] In the nonvolatile semiconductor memory device 51, an insulating layer IS2 is provided on a semiconductor substrate S1, and a fin S3 extending in the y direction is provided on the insulating layer IS2. The insulating layer IS2 is composed of a layered insulating layer 52 formed on the semiconductor substrate S1 and a strip-shaped lower insulating layer 53. The lower insulating layer 53 is made of the same insulating material as the layered insulating layer 52 and is formed integrally with the layered insulating layer 52. The lower insulating layer 53 is extended in the y direction, and a fin S3 is provided on its surface. In addition, in this embodiment, in accordance with the embodiment, Fig.19 The S3 shown is called a "fin", but is different from the Figure 1 The fin S2 shown in FIG. 1 represents a fin shape connected to the semiconductor substrate S2 . In fact, the fin S3 is arranged in a horizontal column shape or a beam shape extending along the y direction.
[0247] The fin S3 is formed of a semiconductor material and is extended along the lower insulating layer 53. In the embodiment in which the lower insulating layer 53 is provided, the width dimension of the fin S3 in a direction (x direction) perpendicular to the length direction (y direction) and the height direction (z direction) of the extended provision is set to be less than the width dimension of the lower insulating layer 53, and the fin S3 is provided only on the lower insulating layer 53.
[0248] In addition, although Fig.19 5. However, an upper insulating layer 54 (using the upper insulating layer 54) is provided on the upper surface of the fin S3 where the memory cell MC4 is provided. Fig. 20A 21 will be described later). Thus, the fin S3 has a structure in which the upper surface and the lower surface are sandwiched by the insulating layer.
[0249] In the fin S3, a drain region 12a is provided on the surface adjacent to the first selection gate structure 3, and a source region 12b is provided on the surface adjacent to the second selection gate structure 4. In this embodiment, the drain region 12a and the source region 12b are formed on the surface of the fin S3 in such a manner that the upper insulating layer 54 is not provided on the upper surface of the fin S3 where the memory cell MC4 is not provided.
[0250] Here, with Figure 3A The corresponding parts are marked with the same reference numerals. Fig. 20A It is shown Fig.19 A cross-sectional view of the cross-sectional structure of the A-A' portion. Fig. 20A As shown, in the fin S3, an upper insulating layer 54 is provided on the surface of a region where the memory gate structure 2, the first select gate structure 3, the second select gate structure 4, the sidewall spacers 6, 7 and the sidewalls 22a, 22b are provided.
[0251] As described above, in the fin S3 in the area where the memory gate structure 2, the first selection gate structure 3, the second selection gate structure 4, the sidewall spacers 6, 7 and the sidewall portions 22a, 22b are provided, the fin S3 has a structure in which the fin S3 is sandwiched by the insulating layer IS2 (the lower insulating layer 53 in this embodiment) and the upper insulating layer 54.
[0252] and Figure 3B The corresponding parts are marked with the same reference numerals. Fig. 20B It is shown Fig.19 The cross-sectional view of the cross-sectional structure of the B-B' portion. Fig. 20B As shown, in the region where the fin S3 is not formed, the memory gate structure 2, the first selection gate structure 3, the second selection gate structure 4, the sidewall spacers 6, 7 and the sidewalls 22a, 22b are formed on the layered insulating layer 52 of the insulating layer IS2 covering the semiconductor substrate S1.
[0253] and Figure 4A The corresponding parts are marked with the same reference numerals. Fig.21A It is shown Fig.19 In this embodiment, as Fig.21A As shown, the upper and lower surfaces of the fin S3 are sandwiched by the upper insulating layer 54 and the insulating layer IS2 (the lower insulating layer 53 in this embodiment), and the side of the fin S3 is surrounded by the memory gate MG, so a channel layer can be formed on the side of the fin S3 in the memory transistor MT. In addition, since the upper and lower surfaces of the fin S3 are sandwiched by the upper insulating layer 54 and the insulating layer IS2 (the lower insulating layer 53 in this embodiment), the memory transistor MT becomes a fully depleted transistor. And, in this embodiment, when the height of the fin S3 is H1fin, it is also formed in a manner that satisfies the relationship Hfin>Wfin.
[0254] In this embodiment, if Fig.21A As shown, the distance H7 from the semiconductor substrate S1 to the lower surface position of the fin S3 is set to be greater than the distance H1 from the semiconductor substrate S1 in the area where the fin S3 is not provided to the lower surface position of the memory gate MG. As a result, the lower surface position of the memory gate can be lower than the lower surface position of the fin S3, and the side of the fin S3 can be completely covered by the memory gate MG, thereby preventing leakage current from flowing along the side area of the fin S3 and between the drain region 12a and the source region 12b to cause malfunction.
[0255] The difference between the distance H7 from the semiconductor substrate S1 to the lower surface of the fin S3 and the distance H1 from the semiconductor substrate S1 to the lower surface of the memory gate MG is preferably not less than 1 / 2 of the width Wfin of the fin S3. The structure without the lower insulating layer 53 will be described below.
[0256] and Figure 4B The corresponding parts are marked with the same reference numerals. Fig.21B It is shown Fig.19 The cross-sectional view of the cross-sectional structure of the D-D' portion. The upper and lower surfaces of the fin S3 are sandwiched by the upper insulating layer 54 and the insulating layer IS2 (the lower insulating layer 53 in this embodiment), and the side of the fin S3 is surrounded by the first selection gate DG, so a channel layer can be formed on the side of the fin S3 in the first selection transistor T1. In addition, since the upper and lower surfaces of the fin S3 are sandwiched by the upper insulating layer 54 and the insulating layer IS2 (the lower insulating layer 53 in this embodiment) and insulated from the semiconductor substrate S1, the first selection transistor T1 becomes a fully depleted transistor.
[0257] In this embodiment, if Fig.21BAs shown, the distance H7 from the semiconductor substrate S1 to the lower surface position of the fin S3 is set to be greater than the distance H2 from the semiconductor substrate S1 in the area where the fin S3 is not provided to the lower surface position of the first selection gate DG. As a result, the lower surface position of the first selection gate can be lower than the lower surface position of the fin S3, and the side of the fin S3 is completely covered by the first selection gate DG, thereby preventing leakage current from flowing along the side area of the fin S3.
[0258] Furthermore, the second selection gate structure 4 also has a structure provided on the layered insulating layer 42 in a manner straddling the fin portion S3, and has the same Fig.21B The first selection gate structure 3 shown has the same structure, so the description thereof is omitted here.
[0259] The difference between the distance H7 from the semiconductor substrate S1 to the lower surface of the fin S3 and the distance H2 from the semiconductor substrate S1 to the lower surface of the first selection gate DG is preferably greater than 1 / 2 of the width Wfin of the fin S3. The structure without the lower insulating layer 53 will be described below.
[0260] The memory cell MC4 according to this embodiment is also the same as the above-mentioned "(3) Circuit structure of nonvolatile semiconductor memory device" and "(4) Voltages during various operations of nonvolatile semiconductor memory device", so their description is omitted.
[0261] As a method for manufacturing the memory cell 4 in which the fin S3 is provided on the insulating layer IS1, Figure 8 Before the fin formation process shown, a semiconductor substrate, an insulating layer, a semiconductor layer as a fin, and an insulating layer as an upper insulating layer are sequentially stacked, and in the fin formation process, the insulating layer as a patterned upper insulating layer is preferably used as a hard mask to process a portion of the semiconductor layer and the insulating layer as the fin. In addition, a semiconductor substrate, an insulating layer, and a semiconductor layer as a fin may also use an insulating silicon (SOI) substrate.
[0262] (10-2) Function and Effect
[0263] Through the above-mentioned structure, this memory cell MC4 also prevents the writing of data by blocking the injection of charges into the charge storage layer EC through the same principle as the above-mentioned embodiment, and suppresses the occurrence of interference by reducing the potential difference between the memory gate MG and the fin S3, and further reducing the electric field of the depletion layer. Therefore, miniaturization can be achieved while suppressing the occurrence of interference.
[0264] In addition, in the memory cell MC4, the upper surface and the lower surface of the fin S3 are sandwiched by the upper insulating layer 54 and the insulating layer IS2 (the lower insulating layer 53 in this embodiment), so that the memory transistor MT, the first selection transistor T1 and the second selection transistor T2 can be made into fully depleted transistors. In this case, since the fin S3 is insulated from the semiconductor substrate S1, the potential of the fin S3 is independent of the semiconductor substrate S1, and the potential difference between the memory gate MG and the fin S3 can be made smaller, and the electric field of the depletion layer can also be made smaller, so compared with the above-mentioned fin type (fin S2), the occurrence of interference can be more controlled.
[0265] In addition, in the memory cell MC4, the lower surface positions of the memory gate MG, the first selection gate DG and the second selection gate SG can be made lower than the lower surface position of the fin S3, and the side surfaces of the fin S3 are completely covered by the memory gate MG, the first selection gate DG and the second selection gate SG, thereby preventing leakage current from flowing along the side areas of the fin S3.
[0266] In addition, in the memory cell MC4, by providing an upper insulating layer 54 covering the upper surface of the fin S3, for example, during a data writing operation, the upper insulating layer 54 can prevent the electric field from being concentrated around the sharp corners where the upper surface of the fin S3 intersects with the side surface. Thus, it is possible to prevent a write failure caused by the electric field only being concentrated around the corners and the inability to inject charges into the charge storage layer film along the fin sidewalls during writing.
[0267] (10-3) Fin according to other embodiments
[0268] Furthermore, in the above-mentioned embodiment, the case is described in which the distance H7 from the semiconductor substrate S1 to the lower surface of the fin S3 is greater than the distance H1 from the semiconductor substrate S1 in the area where the fin S3 is not provided to the lower surface of the memory gate MG (H7>H1), but the present invention is not limited to this. It can also be that the distance H7 from the semiconductor substrate S1 to the lower surface of the fin S3 is less than or equal to the distance H1 from the semiconductor substrate S1 in the area where the fin S3 is not provided to the lower surface of the memory gate MG (H7≦H1).
[0269] For example, in a structure in which a lower insulating layer 53 is provided on a layered insulating layer 52, a distance H7 from the semiconductor substrate S1 to the lower surface of the fin S3 may be greater than a distance H5 from the semiconductor substrate S1 to the layered insulating layer 52, and less than a distance H6 from the semiconductor substrate S1 to the lower surface of the fin S3. Fig.21A A distance H1 (H5 < H7 < H1) from the semiconductor substrate S1 to the lower surface of the memory gate MG.
[0270] In addition, in the above-mentioned embodiment, the lower insulating layer 53 protruding from the layered insulating layer 52 is provided on the layered insulating layer 52, and the layered insulating layer 52, the lower insulating layer 53, the fin S3 and the upper insulating layer 54 are sequentially stacked from the semiconductor substrate S1. However, the present invention is not limited thereto, and the lower insulating layer 53 may not be provided, but the layered insulating layer 52, the fin S3 and the upper insulating layer 54 may be sequentially stacked from the semiconductor substrate S1. In this case, the lower surface position of the fin S3 may be made to be in the same plane as the surface position of the layered insulating layer 52.
[0271] Furthermore, in the above-mentioned embodiment, a structure is described in which the upper insulating layer 54 is provided on the fin S3, and the insulating layer IS2, the fin S3 and the upper insulating layer 53 are stacked in sequence from the semiconductor substrate S1. However, the present invention is not limited to this, and a structure in which the upper insulating layer 54 is not provided and the insulating layer IS2 and the fin S3 are stacked in sequence from the semiconductor substrate S1 can also be made.
[0272] And, in Fig.16 and Fig.19 In the embodiment of the present invention, the memory gate MG, the first selection gate DG, and the second selection gate SG are formed of a metal material, but the present invention is not limited thereto. The memory gate MG, the first selection gate DG, and the second selection gate SG may be formed of a conductive material other than a metal material (e.g., polysilicon) as in the above-mentioned FIG. 12 . In this case, Fig.16 and Fig.19 In the manufacturing process of the memory cells MC3 and MC4 shown, for example, the layered electrode forming layer is etched back to form the sidewall-shaped first and second selection gates DG1 and SG1 along the sidewall spacers 6 and 7 of the sidewalls of the memory gate structure 2a.
[0273] Description of Reference Numerals
[0274] 1, 31, 41, 51: Nonvolatile semiconductor memory devices
[0275] MC, MC11, MC12, ..., MC1n, MC21, MC22, ..., MC2n, MCm1, MCm2, ..., MCmn, MC1, MC2, MC3, MC4: memory cell
[0276] 2.2a: Memory gate structure
[0277] 3. 3a: First selection gate structure
[0278] 4.4a: Second selection gate structure
[0279] 12a, 40a: Drain region
[0280] 12b, 40b: Source region
[0281] 6, 7: Side wall partition
[0282] 13: Lower memory gate insulation film
[0283] 14, 14a: Upper memory gate insulating film
[0284] EC: Charge storage layer
[0285] 17a, 18a: first selection gate insulating film
[0286] 17b, 18b: second selection gate insulating film
[0287] S1: Semiconductor substrate
[0288] S2, S3: Fins
[0289] IS, IS1, IS2: Insulation layer
[0290] 42, 52: Layered insulation layer
[0291] 43: Convex insulation layer
[0292] 53: Lower insulation layer
[0293] 54: Upper insulation layer
Claims
1. A memory cell, characterized in that: include: a semiconductor substrate, covered by an insulating layer; a fin portion formed on the insulating layer; A memory gate structure, which is stacked with a lower memory gate insulating film, a charge storage layer, an upper memory gate insulating film and a memory gate, and is formed on the insulating layer in a manner spanning the fin; A first selection gate structure having a first selection gate disposed on a first selection gate insulating film, and formed on the insulating layer along a sidewall spacer formed on one sidewall of the memory gate structure in a manner spanning the fin; A second selection gate structure having a second selection gate disposed on the second selection gate insulating film and formed on the insulating layer along another sidewall spacer formed on the other sidewall of the memory gate structure in a manner spanning the fin; a drain region provided on a surface of the fin adjacent to the first select gate structure in a manner insulated from the first select gate and electrically connected to a bit line; a source region provided on a surface of the fin adjacent to the second selection gate structure in a manner insulated from the second selection gate and electrically connected to a source line; The first selection gate structure, the memory gate structure, and the second selection gate structure are disposed between the drain region and the source region. An upper insulating layer is provided on the upper surface of the fin portion corresponding to the region where the memory gate structure, the side wall spacer, the first selection gate structure, the other side wall spacer, and the second selection gate structure are provided. The insulating layer comprises: a layered insulating layer disposed on the semiconductor substrate; a lower insulating layer, which is disposed on the layered insulating layer and is formed integrally with the layered insulating layer and is made of the same material as the layered insulating layer; a fin is disposed on the surface of the lower insulating layer; and the layered insulating layer, the lower insulating layer, the fin and the upper insulating layer are sequentially stacked from the semiconductor substrate. The width dimension of the layered insulating layer in a direction perpendicular to a length direction in which the fin extends and a height direction of the fin is greater than a width dimension of the lower insulating layer.
2. The memory unit according to claim 1, characterized in that A distance from the surface of the semiconductor substrate to the lower surface of the fin is greater than a distance from the surface of the semiconductor substrate to the surface of the layered insulating layer in a region where the fin is not provided.
3. The memory unit according to claim 1 or 2, characterized in that The memory gate, the first select gate, and the second select gate include a metal material.
4. The memory unit according to claim 1 or 2, characterized in that: The first selection gate is formed in a sidewall shape along the one sidewall spacer, and the second selection gate is formed in a sidewall shape along the other sidewall spacer.
5. The memory unit according to claim 1 or 2, characterized in that: When the height of the fin, which is the distance from the surface of the lower insulating layer to the lower surface of the upper insulating layer, is represented by H1fin, and the width of the fin across the fin direction of the first selection gate structure and the second selection gate structure is represented by Wfin, Hfin>Wfin.
6. A nonvolatile semiconductor storage device, characterized in that: The memory cells connected to the memory gate lines are arranged in a matrix. The memory cell is the memory cell according to claim 1 or 2, and a plurality of the memory cells arranged in a matrix share the memory gate line.
