Memory array circuit
By forming bit lines and complementary bit lines under the surface of the semiconductor substrate and connecting transistors with vertical interconnect lines, the problem of improving the size and performance of integrated circuits in traditional technology is solved, and a more compact structure and higher performance is achieved.
Patent Information
- Application Number
- CN202410665656.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2024-05-24
- Filing Date
- 2024-05-27
- Publication Date
- 2024-11-26
AI Technical Summary
The prior art is difficult to effectively reduce the size of integrated circuits and improve performance, especially in transistor connections and signal transmission. Traditional dynamic random access memory structures have problems of area waste and noise increase.
Under-surface interconnection technology is adopted to achieve a more compact structure and higher performance by forming bit lines and complementary bit lines below the original surface of the semiconductor substrate and connecting transistors with vertical interconnections, reducing the use of metal lines.
It realizes a smaller integrated circuit size, reduces noise and power consumption, and improves the AC performance and signal transmission efficiency of the circuit.
Smart Images

Figure CN119028398A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a memory array circuit, and in particular to a memory array circuit which can provide subsurface interconnects or subsurface interconnects buried in a silicon substrate, thereby having the advantages of compact structure, smaller area, higher performance, lower complexity, etc. Background Art
[0002] In the most advanced integrated circuits in the prior art, there are many transistors connected by conductive interconnects (such as metal lines, polysilicon lines, etc.) to facilitate signal transmission between the gate region, source region and drain region of these transistors. All of these conductive interconnects rely on many contact holes and connecting plugs to connect to the gate region, the source region and the drain region respectively, which poses a significant challenge and difficulty to the chip design goals of reducing the area, power consumption and noise of these transistors and improving the performance of integrated circuits, especially when the size of components needs to be reduced to meet Moore's Law, the number of integrated circuits on the chip must be significantly reduced. To give an example of area loss: compared with the size of the contact hole used to connect the metal line to the source region or the drain region, a larger source region or drain region diffusion area must be designed so that the inevitable photolithographic misalignment due to the limitation of the lithographic tool should not cause the contact hole to be formed outside the edge of the source region or the drain region. In this way, the diffusion area of the transistor will inevitably increase, thereby increasing the chip area, resulting in an increase in capacitance and a serious impact on the AC performance of the circuit, consuming more power and increasing greater noise.
[0003] In addition, in the structure of traditional dynamic random access memory, there are multiple sensor amplifiers connected to the dynamic random access memory cell array through metal lines. The design rules of these metal lines set above the silicon substrate and the spacing between the metal lines will limit the shrinkability of the dynamic random access memory size. Therefore, how to introduce better self-aligned contact structures and technologies to connect transistors to their first interconnect (metal) layer with the smallest surface area to transmit and receive signals is a key challenge to further effectively reduce the size of integrated circuits and improve performance. Summary of the invention
[0004] An embodiment of the present invention provides a memory array circuit. The memory array circuit includes a semiconductor substrate, a bit line, a complementary bit line, and a bit line sense amplifier. The semiconductor substrate has an original surface. The bit line sense amplifier is connected to the bit line and the complementary bit line, wherein the bit line sense amplifier includes a first plurality of transistors and a first group of connecting wires. Each transistor includes a gate node, a first conductive node, and a second conductive node. The first group of connecting wires connects the first conductive nodes of the first plurality of transistors to the bit line and the complementary bit line. The first group of connecting wires is located below the original surface of the semiconductor substrate.
[0005] In one embodiment of the present invention, the first set of connection lines is located below the original surface of the semiconductor substrate, and at least a portion of the first set of connection lines is isolated from the semiconductor substrate.
[0006] In an embodiment of the present invention, the bit line and the complementary bit line are located below an original surface of the semiconductor substrate and isolated from the semiconductor substrate.
[0007] In one embodiment of the present invention, the memory array circuit further comprises a second set of connection lines connecting the gate nodes of the first plurality of transistors to the bit lines and the complementary bit lines, wherein the second set of connection lines is located above the original surface of the semiconductor substrate.
[0008] In one embodiment of the present invention, each connection line in the first set of connection lines extends downward from the original surface to the bit line or the complementary bit line, and the bit line or the complementary bit line extends laterally.
[0009] In one embodiment of the present invention, the memory array circuit further includes an equalization circuit, wherein a third group of connection lines connecting the equalization circuit to the bit line and the complementary bit line is located below the original surface of the semiconductor substrate, and each connection line in the third group of connection lines extends downward from the original surface to the bit line or the complementary bit line.
[0010] In one embodiment of the present invention, the bit line sense amplifier further includes a local data line connected to a bit line switch and a local complementary data line connected to a complementary bit line switch, wherein the local data line and the local complementary data line lines are located below the original surface of the semiconductor substrate and isolated from the semiconductor substrate.
[0011] In one embodiment of the present invention, a fourth set of connecting lines connecting the bit line switch and the complementary bit line switch to the bit line and the complementary bit line is located below the original surface of the semiconductor substrate, and each connecting line in the fourth set of connecting lines extends downward from the original surface to the bit line or the complementary bit line.
[0012] In one embodiment of the present invention, the bit line is located below the original surface of the semiconductor substrate, and the distance between the original surface of the semiconductor substrate and the local data line is smaller than the distance between the original surface of the semiconductor substrate and the bit line.
[0013] In one embodiment of the present invention, the first plurality of transistors include two P-type metal-oxide-semiconductor (PMOS) transistors and two N-type metal-oxide-semiconductor (NMOS) transistors, a fifth group of connecting lines connecting the second conductive nodes of the two PMOS transistors are located below the original surface of the semiconductor substrate, and a sixth group of connecting lines connecting the second conductive nodes of the two NMOS transistors are located below the original surface of the semiconductor substrate.
[0014] In one embodiment of the present invention, the fifth group of connection lines and the sixth group of connection lines extend laterally.
[0015] An embodiment of the present invention provides a memory array circuit. The memory array circuit includes a semiconductor substrate, a bit line and a complementary bit line and a bit line sense amplifier. The semiconductor substrate has an original surface. The bit line and the complementary bit line are located below the original surface of the semiconductor substrate. The bit line sense amplifier is connected to the bit line and the complementary bit line, wherein the bit line sense amplifier includes a first plurality of transistors, a group of bit line switches, a local data line and a local complementary data line, and an external data line and an external complementary data line. The first plurality of transistors are coupled to the bit line and the complementary bit line. The group of bit line switches is coupled to the bit line and the complementary bit line. The local data line and the local complementary data line are coupled to the group of bit line switches. The external data line and the external complementary data line are coupled to the group of bit line switches through the local data line and the local complementary data line. The local data line and the local complementary data line are located below the original surface of the semiconductor substrate.
[0016] In one embodiment of the present invention, the first plurality of transistors and the set of bit line switches are disposed between a horizontal pitch between the bit line and the local data line.
[0017] In one embodiment of the present invention, the memory array circuit further comprises an equalization circuit having a second plurality of transistors, and the second plurality of transistors are disposed between the horizontal spacing between the bit lines and the local data lines.
[0018] In one embodiment of the present invention, the horizontal spacing between the bit line and the local data line is less than 3.5λ, and λ is a technology node of the memory array circuit.
[0019] In one embodiment of the present invention, the technology node λ is lower than 30 nanometers (nm).
[0020] In one embodiment of the present invention, the length of the bit line is substantially the same as the length of the complementary bit line.
[0021] In one embodiment of the present invention, a first set of connecting lines connecting the first plurality of transistors to the bit lines and the complementary bit lines is located below the original surface of the semiconductor substrate, and a second set of connecting lines connecting the first plurality of transistors to the bit lines and the complementary bit lines is located above the original surface of the semiconductor substrate.
[0022] In an embodiment of the present invention, a distance between the original surface of the semiconductor substrate and the local data line is shorter than a distance between the original surface of the semiconductor substrate and the bit line. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] Figure 1A The present invention is a flowchart of a method for manufacturing a memory array circuit according to an embodiment of the present invention.
[0024] Figure 1B , Figure 1C , Figure 1D , Figure 1E , Figure 1F , Figure 1G Yes Description Figure 1A Schematic diagram of .
[0025] Figure 2 It is a schematic diagram illustrating a cross-sectional view and a top view along the X direction after a liner oxide layer is grown, a liner nitride layer is deposited, and a trench is formed.
[0026] Figure 3 is a schematic diagram illustrating deposition and etching back of a first nitride layer to form a first nitride spacer layer, and deposition and polishing of a spin-on dielectric.
[0027] Figure 4 It is a schematic diagram for explaining the steps of etching away the upper edge of the first nitride spacer layer, spinning on a dielectric, generating a first oxide layer, and depositing a metal layer.
[0028] Figure 5 is a schematic diagram illustrating etching back a metal layer to form a subsurface bit line.
[0029] Figure 6 is a schematic diagram illustrating deposition of a second oxide layer in the trench and planarization by the chemical mechanical polishing technique.
[0030] Figure 7 is a schematic diagram illustrating depositing a third oxide layer, a second nitride layer, and a patterned photoresist layer, and then etching away unnecessary portions of the third oxide layer, the second nitride layer, and the patterned photoresist layer.
[0031] Figure 8 It is a schematic diagram illustrating forming a U-shaped recess, forming a high dielectric constant insulating layer in the U-shaped recess, depositing and etching back a gate material to form a word line and a gate structure of the access transistor, depositing and etching back a third nitride layer and a fourth oxide layer, and etching away the second nitride layer and the third oxide layer.
[0032] Fig. 9 It is a schematic diagram for explaining the removal of the pad nitride layer, etching back the second oxide layer, depositing and anisotropically etching the fourth nitride layer, the fifth oxide layer and the fifth nitride layer, and depositing the sixth oxide layer.
[0033] Fig.10 It is a schematic diagram illustrating the deposition and polishing of a spin-on dielectric, the deposition of a photoresist layer, the etching back of the spin-on dielectric to be removed, the etching away of the exposed sixth oxide layer, the pad oxide layer, and the silicon material corresponding to the horizontal silicon surface HSS-1 / 3 to produce a hole hole-1 / 3, the thermal generation of a seventh oxide layer, and the deposition and etching back of the spin-on dielectric.
[0034] Fig.11 It is a schematic diagram illustrating the removal of the exposed spin-on dielectric, the exposed sixth oxide layer and the exposed pad oxide layer, and then digging out the silicon material corresponding to the horizontal silicon surface HSS-1 / 2 to generate a hole hole-1 / 2, and thermally generating an eighth oxide layer.
[0035] Fig.12 It is a schematic diagram illustrating a cross-sectional view of the memory array circuit along the Y2 direction, wherein the Y2 direction extends along the center of the hole Hole-1 / 2 and is perpendicular to the X direction.
[0036] Fig.13 It is a schematic diagram illustrating removing the first nitride spacer at the lower edge to expose the sidewall of the subsurface bit line, and depositing an n+ heavily doped polysilicon film in the hole hole-1 / 2 to contact the sidewall of the subsurface bit line. Fig.14 is a schematic diagram illustrating etching back the upper portion of the eighth oxide layer to expose the silicon material corresponding to the hole hole-1 / 2.
[0037] Fig.15 It is a schematic diagram for explaining the lateral generation of an n+ polysilicon layer and the thermal generation of a ninth oxide layer based on the exposed silicon material.
[0038] Fig.16It is a schematic diagram of realizing the vertical connection (bridge) of the access transistor and the drain region and how they are connected to the subsurface interconnection line disclosed in another embodiment of the present invention.
[0039] Fig.17 is a schematic diagram illustrating the formation of the drain region and the source region of the access transistor.
[0040] Fig.18 is a schematic diagram illustrating the formation of two different subsurface interconnections at two different vertical heights in a pipeline and the separation of the two from each other by an isolation material.
[0041] Fig.19 is a schematic diagram illustrating waveforms of related signals of a conventional dynamic random access memory cell during a write operation.
[0042] Fig. 20 FIG. 5 is a schematic diagram illustrating a circuit for selectively coupling the sense amplifier to two separate voltages during a write operation of the DRAM cell.
[0043] Fig.21 is a schematic diagram illustrating the most commonly used design of the dynamic random access memory cell.
[0044] Fig. 22 is a schematic diagram illustrating a conventional sensor amplifier and an equalization circuit design circuit.
[0045] Fig.23 and Fig.24 is a schematic diagram illustrating a conventional bit line sense amplifier design for open-bit line's of a dynamic random access memory.
[0046] Fig.25 is a schematic diagram illustrating the coupling relationship between the active region, the connecting pin and the metal 0.
[0047] Fig.26 is a schematic diagram illustrating a bit line within a bit line sense amplifier.
[0048] Fig. 27 is a schematic diagram illustrating that part of the connecting wires in a conventional bit line sense amplifier are replaced by subsurface connectors.
[0049] Fig.28A and Fig.28B It is a schematic diagram illustrating that the subsurface interconnect technology is applied in the layout of the bit line sense amplifier to reduce the width and length of the bit line sense amplifier.
[0050] Fig.29 , Fig. 30A and Fig. 30B is a schematic diagram illustrating the layout of a bit line sense amplifier having subsurface connectors and elevated conductor post connections.
[0051] Fig.31 and Fig.32 The present invention is a schematic diagram illustrating that a metal 0 connection in a P-type metal oxide semiconductor transistor or an N-type metal oxide semiconductor transistor in a bit line sense amplifier is replaced by a subsurface connector connection. Fig.33A , Fig.33B , Fig.33C , Fig.33D , Fig.33E Yes Description Fig. 30B Schematic diagram of the step-by-step partial layout.
[0052] The reference numerals are described as follows:
[0053] 11 Access transistor
[0054] 12 Storage capacitor
[0055] 80 sensor amplifier
[0056] 70 Data Line Sense Amplifier
[0057] 75 memory array
[0058] 202 p-type substrate
[0059] 204 pad oxide layer
[0060] 206 liner nitride layer
[0061] 208, HSS-1 / 2, HSS-1 / 3, HSS horizontal silicon surface
[0062] 210 Groove
[0063] 214 oxide layer
[0064] 304, 1902, 2202 spin-on dielectrics
[0065] 306, 1106 photoresist layer
[0066] 502 first oxide layer
[0067] 602 metal layer
[0068] 902, UGBL, UGBL0B, UGBL1B surface lower bit line
[0069] 1002, CVD-STI-Oxide2 second oxide layer
[0070] 1102 Third Oxide Layer
[0071] 1104 Second Nitride Layer
[0072] 1302 U-shaped channel
[0073] 1304 high dielectric constant insulation layer
[0074] 1306 Gate Material
[0075] 1402 third nitride layer
[0076] 1404 Fourth Oxide Layer
[0077] 1602 Fourth Nitride Layer
[0078] 1604 Fifth Oxide Layer
[0079] 1606 Fifth Nitride Layer
[0080] 1802 Sixth Oxide Layer
[0081] 2203, EQBL equalization circuit
[0082] 2102 Seventh Oxide Layer
[0083] 2402 Eighth Oxide Layer
[0084] 2802 n+ heavily doped polysilicon film
[0085] 3202 n+ polysilicon layer
[0086] 3302 Ninth Oxide Layer
[0087] AQ1, AQ2, AQ3 access transistors
[0088] BL, BL0, BL1 bit lines
[0089] BLB, BL0B, BL1B complementary bit lines
[0090] BS100, BS1, BS2, BS0, BS position switch
[0091] BLSA0, BLSA1, BLSAn-1, BLSAn, bit line sense amplifier
[0092] BLSA0_new, BLSA1_new
[0093] DI data input circuit Drain-1 first drain
[0094] Drain-2 Second drain
[0095] DL data cable
[0096] DLB complementary data line
[0097] EN1, EN2, SLNS, eqpl control signals
[0098] E1 first end
[0099] E2 Second Terminal
[0100] ECoP1, ECoP2, ECoP elevated conductor columns
[0101] GIO global input / output path
[0102] Gate
[0103] Hole-1 / 2, Hole-1 / 3
[0104] LSLP, LSLN, 1~23, 1a~23a connection line
[0105] LDL, LDLB local data line
[0106] N1, N2, NSA N-type metal oxide semiconductor transistor N-WELL N well
[0107] NMOS-AA Active region of N-type metal oxide semiconductor transistor
[0108] NMOS-gate N-type metal oxide semiconductor transistor gate oxide-7plug seventh oxide plug
[0109] P1, P2, PSA P-type metal oxide semiconductor transistor PMOS-AA P-type metal oxide semiconductor transistor active area
[0110] PMOS-gate The gate of the P-type metal oxide semiconductor transistor POLY polysilicon
[0111] P20, P40, N10, N20, N50, N54, transistors
[0112] N55, N30, N40, SLP, SLNSource-1 first source
[0113] STI shallow trench isolation
[0114] Source-3 The third source
[0115] UGIL1, UGIL2, UGIL subsurface interconnects
[0116] UGC1, UGC2, UGC subsurface connectors
[0117] Vpl common voltage
[0118] VCCSA, VSS supply voltage
[0119] VSSSA, VBL voltage
[0120] Tungsten plug
[0121] WL word line
[0122] Wordline-1 First wordline
[0123] Wordline-2 Second wordline
[0124] Wordline-3 The third wordline
[0125] XIO write data
[0126] Steps 10-50, 102-156 DETAILED DESCRIPTION
[0127] In order to illustrate that the present invention covers such a wide range of integrated circuits including logic or digital, analog, memory, etc., the power connection circuit is first selected as the focus in the following description of the present invention to make the technical features of the present invention easier to grasp.
[0128] Please refer to Figure 1A , Figure 1B , Figure 1C , Figure 1D , Figure 1E , Figure 1F , Figure 1G ,in Figure 1A The present invention is a flowchart of a method for manufacturing a memory array circuit according to an embodiment of the present invention. Figure 1A Include:
[0129] Step 10: Start;
[0130] Step 15: On the basis of the substrate (i.e., the p-type substrate 202, and the present invention takes the p-type substrate 202 as a silicon substrate as an example), define the active region of the memory array circuit and form shallow trench isolation (STI);
[0131] Step 20: forming an asymmetric spacer layer along the sidewall of the active region;
[0132] Step 25: forming an underground interconnection line (an underground bitline is used as an example in the present invention) between the asymmetric spacer layers and below the horizontal surface of the substrate;
[0133] Step 30: forming a cell line and a gate of an access transistor of the memory array circuit;
[0134] Step 35: defining the drain region and the source region of the access transistor of the memory array circuit;
[0135] Step 40: forming a connection between the subsurface bit line and the drain region of the access transistor;
[0136] Step 45: forming the drain region and the source region;
[0137] Step 50: End.
[0138] Please refer to Figure 1B and Figure 2 , step 15 comprises:
[0139] Step 102: Depositing a pad oxide layer 204 and a pad nitride layer 206 on a horizontal surface of the substrate (the present invention takes the p-type substrate 202 as a silicon substrate as an example, so the horizontal surface is a horizontal silicon surface (horizontal silicon surface 208), and the horizontal silicon surface 208 is also called HSS) Figure 2 );
[0140] Step 104: define the active region of the memory array circuit, and remove the substrate material (ie, silicon material) outside the active region corresponding to the horizontal silicon surface 208 to form a trench 210 ( Figure 2 );
[0141] Step 106: Deposit and etch back the oxide layer 214 in the trench 210 to form the shallow trench isolation (STI) below the horizontal silicon surface 208. Figure 2 ).
[0142] Please refer to Figure 1C , Figure 3 , Figure 4 , step 20 comprises:
[0143] Step 108: depositing and etching back the first nitride layer to form a first nitride spacer, and depositing a spin-on dielectric (SOD) 304 in the trench 210 and planarizing the SOD 304 by chemical mechanical polishing (CMP) technology ( Figure 3 );
[0144] Step 110: Etch away the upper edge of the first nitride spacer layer, the first nitride spacer layer and the spin-on dielectric 304 ( Figure 4 );
[0145] Step 112: Generate a first oxide layer 502 (eg, thermally generated) and deposit a metal layer 602 in the trench 210 and planarize it by the chemical mechanical polishing technique. Figure 4 ).
[0146] Please refer to Figure 1D , Figure 5 , Figure 6 , step 25 comprises:
[0147] Step 114: Use photolithographic technique to form a plurality of interconnection (ie metal layer 602) lines ( Figure 5 );
[0148] Step 116: Etch back the metal layer 602 (the plurality of interconnects) to form underground bitlines (UGBL) 902 ( Figure 5 );
[0149] Step 118: Deposit the second oxide layer 1002 in the trench 210 and planarize it by the chemical mechanical polishing technique. Figure 6 ).
[0150] Please refer to Figure 1E , Figure 7 , Figure 8 , step 30 comprises:
[0151] Step 120: deposit a thick third oxide layer 1102, a thick second nitride layer 1104 and a patterned photoresist layer 1106, and then etch away unnecessary portions of the third oxide layer 1102 and the second nitride layer 1104 ( Figure 7 );
[0152] Step 122: Remove the patterned photoresist layer 1106, the pad nitride layer 206 and the pad oxide layer 204, and expose the horizontal silicon surface 208 ( Figure 7 );
[0153] Step 124: Etching the exposed horizontal silicon surface 208 to form a U-shaped recess, forming a high-k insulating layer 1304 in the U-shaped recess, and depositing and etching back a gate material 1306 (e.g., tungsten) to form a gate line and gate structure of the access transistor, wherein the access transistor can be referred to as a U-shaped transistor ( Figure 8 );
[0154] Step 126: deposit and etch back the third nitride layer 1402, then deposit the fourth oxide layer 1404, and then etch back or planarize the fourth oxide layer 1404 ( Figure 8 );
[0155] Step 128: Etch away the second nitride layer 1104 and the third oxide layer 1102 ( Figure 8 ).
[0156] Please refer to Figure 1F , Fig. 9 , Fig.10 , step 35 comprises:
[0157] Step 130: Remove the pad nitride layer 206 and etch back the second oxide layer 1002 to the top of the pad oxide layer 204 ( Fig. 9 );
[0158] Step 132: Deposit and anisotropically etch the fourth nitride layer 1602, the fifth oxide layer 1604 and the fifth nitride layer 1606 ( Fig. 9 );
[0159] Step 134: Deposit the sixth oxide layer 1802 ( Fig. 9 );
[0160] Step 136: Deposit and then planarize the spin-on dielectric 1902 by the chemical mechanical planarization technique, deposit a photoresist layer, and then etch back the spin-on dielectric 1902 to be removed ( Fig.10 );
[0161] Step 138: Etch away the exposed sixth oxide layer 1802, the pad oxide layer 204, and the silicon material corresponding to the horizontal silicon surface HSS-1 / 3 to form a hole hole-1 / 3 ( Fig.10 );
[0162] Step 140: Thermally generate the seventh oxide layer 2102 ( Fig.10 );
[0163] Step 142: Deposit a spin-on dielectric 2202 on the seventh oxide layer 2102, and then etch back the spin-on dielectric 2202 ( Fig.10 ).
[0164] Please refer to Figure 1G , Fig.11 , Fig.12 , Fig.13 , Fig.14 , Fig.15 , Fig.16 , step 40 comprises:
[0165] Step 144: remove the exposed spin-on dielectric 1902, the exposed sixth oxide layer 1802 and the exposed pad oxide layer 204, and then dig out the silicon material corresponding to the horizontal silicon surface HSS-1 / 2 to form a hole hole-1 / 2 ( Fig.11 );
[0166] Step 146: Thermally generate the eighth oxide layer 2402 ( Fig.11 );
[0167] Step 148: Remove the first nitride spacer at the lower edge to expose the sidewall of the lower bit line (UGBL) ( Fig.13 );
[0168] Step 150: Deposit an n+ heavily doped polysilicon film 2802 in the hole hole-1 / 2 to contact the sidewall of the lower bit line (UGBL) ( Fig.13 );
[0169] Step 152: Etch back the upper portion of the eighth oxide layer 2402 to expose the silicon material corresponding to the hole hole-1 / 2 ( Fig.14 );
[0170] Step 154: Based on the exposed silicon material, an n+ polysilicon layer 3202 is laterally grown to contact the drain region and the n+ polysilicon plug ( Fig.15 );
[0171] Step 156: Thermally generate a ninth oxide layer 3302 on the n+ polysilicon layer 3202 ( Fig.15 ).
[0172] The metal layers used in the above process steps (eg Figure 5 The lower bit line (UGBL) 902 is shown. Fig.10 The material of the bit lines, electrodes of capacitors and / or common electrodes shown in the figure can be tungsten, however, due to the sensitivity of tungsten materials to oxides or oxidation processes, it is best that the tungsten material can be covered by another titanium nitride layer or a suitable layer. Because the characteristics of the tungsten material should be clear to the technicians in the field of the present invention, the present invention does not describe the detailed protection process of the tungsten material, but assumes that the metal layers including the tungsten material are well treated to avoid any oxidation reaction directly thereon. Of course, there are some suitable metal layers that can still be applied to the subsurface bit lines and the bit lines, without being limited to a specific type of metal material that is not suitable for insertion into the integrated circuit process.
[0173] The detailed description of the above manufacturing method is as follows, starting from the p-type substrate 202. In step 102, as Figure 2 As shown in FIG. 5( a ), a pad oxide layer 204 is formed on a horizontal silicon surface 208 (ie, the original surface as described in claim 1 ), and then a pad nitride layer 206 is deposited on the pad oxide layer 204 .
[0174] In step 104, the active region of the memory array circuit may be defined by photolithographic technique, wherein Figure 2 As shown in (a), since the active region of the memory array circuit corresponds to the pad oxide layer 204 and the pad nitride layer 206, the horizontal silicon surface 208 outside the active region pattern will be exposed accordingly. Since the horizontal silicon surface 208 outside the active region pattern is exposed, the silicon material corresponding to the horizontal silicon surface 208 outside the active region pattern can be removed by anisotropic etching technique to produce a trench (or channel) 210, wherein, for example, the trench 210 can be 250 nanometers (nm) deep below the horizontal silicon surface 208.
[0175] In step 106, an oxide layer 214 is deposited to fill the trench 210, and then the oxide layer 214 is etched back to form the shallow trench isolation below the horizontal silicon surface 208 in the trench 210. In addition, Figure 2 (b) corresponds to Figure 2 (a) is a top view, where Figure 2 (a) is along the Figure 2 (b) is a cross-sectional view of the cutting line in the X direction. Figure 2 As shown in (a), for example, the shallow trench isolation has a thickness of about 50 nanometers, and if the trench 210 below the horizontal silicon surface 208 has a depth of 250 nanometers, the top of the shallow trench isolation is about 200 nanometers deep from the horizontal silicon surface 208.
[0176] In step 108, if Figure 3 As shown in (a), the first nitride layer is deposited and the first nitride layer is etched back using the anisotropic etching technology to form the first nitride spacer along the two edges of the trench 210 (that is, the first nitride spacer at the upper edge and the first nitride spacer at the lower edge). In another embodiment of the present invention, the first nitride spacer formed on one of the two edges of the trench 210 can be replaced by silicon oxycarbonitride (SiOCN). Then, as Figure 3 As shown in (a), a spin-on dielectric 304 is deposited on the shallow trench isolation in the trench 210 to fill the vacancy above the shallow trench isolation in the trench 210. The spin-on dielectric 304 is then planarized by the chemical mechanical planarization technique so that the top of the spin-on dielectric 304 is flush with the top of the pad nitride layer 206. In addition, Figure 3 (b) corresponds to Figure 3 (a) is a top view, where Figure 3 (a) is along the Figure 3(b) is a cross-sectional view taken along the cutting line in the Y direction.
[0177] In step 110, if Figure 4 As shown in FIG. 1 , the first nitride spacer layer at the lower edge of the groove 210 is protected by the photoresist layer 306 using the lithography technology, but the first nitride spacer layer at the upper edge of the groove 210 is not protected by the photoresist layer 306. That is, the photoresist layer 306 is deposited on the spin-on dielectric 304 and the liner nitride layer 206 (as shown in FIG. 1 ). Figure 3 After (a), because the photoresist layer 306 on the upper edge first nitride spacer is removed but the photoresist layer 306 on the lower edge first nitride spacer is retained, the lower edge first nitride spacer can be retained but the upper edge first nitride spacer will be removed.
[0178] In step 112, if Figure 4 As shown, the upper edge first nitride spacer layer and the spin-on dielectric 304 not covered by the photoresist layer 306 can be etched away by an etching process. Figure 4 As shown, the photoresist layer 306 and the spin-on dielectric 304 are stripped, wherein the spin-on dielectric 304 has a much higher etching rate than the thermally generated oxide layer and the deposited oxide layer. Then the first oxide layer 502 is thermally generated to form a first oxide spacer, wherein the first oxide spacer covers the upper edge of the trench 210. Figure 4 As shown, step 112 causes an asymmetric spacer (i.e., the first nitride spacer at the lower edge and the first oxide spacer) to be formed at two symmetrical edges of the trench 210 (i.e., the upper edge and the lower edge of the trench 210). For example, the thickness of the first oxide spacer is 4 nanometers and the thickness of the first nitride spacer at the lower edge is 3 nanometers. On the other hand, the asymmetric spacer (i.e., the first oxide spacer and the first nitride spacer at the lower edge) is also formed along the sidewall of the active region. The asymmetric spacer (i.e., Figure 4 The structure of the present invention and the above-mentioned related steps are the first main technical features of the present invention, which are called asymmetric spacers on two symmetrical edges of a trench or a canal (ASoSE).
[0179] In addition, if Figure 4 As shown, a metal layer 602 (e.g., a thin titanium nitride layer plus tungsten) or a conductive material (e.g., doped polysilicon) that needs to withstand subsequent process conditions is deposited to completely fill the trench 210, and is planarized by the chemical mechanical polishing technique so that the top of the metal layer 602 is flush with the top of the pad nitride layer 206.
[0180] Then, in step 114, if Figure 5 As shown in (a), the lithography technique is used to form a plurality of conductive lines (e.g., titanium nitride / tungsten lines). That is, the metal layer 602 is etched back using the lithography technique to leave a metal layer 602 of a reasonable thickness inside the trench 210, thereby forming conductive lines or subsurface bit lines (UGBL) 902, wherein the top of the subsurface bit line 902 is much lower than the horizontal silicon surface HSS (e.g., the thickness of the subsurface bit line 902 is about 40 nm). In addition, as shown in FIG. Figure 5 As shown in (a), the UGBL 902 is located on the top of the STI, and the two sidewalls of the UGBL 902 are defined by the asymmetric spacer (ie, the first oxide spacer and the lower edge first nitride spacer). Figure 5 (a) is along Figure 5 (b) is a cross-sectional view taken along the Y direction.
[0181] In step 118, if Figure 6 (Along as Figure 5 As shown in (b) of the cross-sectional view of the cutting line in the Y direction, the second oxide layer 1002 (also called CVD-STI-oxide2) needs to have a sufficient thickness to fill the trench 210 above the bit line 902 below the surface, and then the second oxide layer 1002 is planarized by the chemical mechanical planarization technology to retain a portion of the second oxide layer 1002, wherein the top of the retained portion of the second oxide layer 1002 is flush with the top of the pad nitride layer 206 and covers the lower edge first nitride spacer and the first oxide spacer. Figure 6 As shown, step 118 can make the subsurface bit line 902 (that is, the interconnection wire) embedded in all insulators (that is, the isolation region) in the trench 210 and confined by all the insulators (the subsurface bit line 902 will then be connected to the drain of the access transistor of the memory array circuit), wherein Figure 6 The structure shown is called a subsurface bit line surrounded by insulator, and the subsurface bit line (UGBL) is the second main technical feature of the present invention.
[0182] The following describes how to form the access transistor and the gate line of the memory array circuit, and the gate structure of the gate line connecting the access transistor. Figure 7As shown in (a), firstly, a third oxide layer 1102, a second nitride layer 1104, and a patterned photoresist layer 1106 are deposited. Then, the unnecessary parts of the third oxide layer 1102 and the second nitride layer 1104 are removed by the lithography technology. In addition, a transistor / word line pattern can be defined by a composite layer composed of the third oxide layer 1102 and the second nitride layer 1104, wherein the composite layer composed of the third oxide layer 1102 and the second nitride layer 1104 is composed of a plurality of strip-shaped third oxide layers 1102 and the second nitride layer 1104 in a direction perpendicular to the active region direction. Therefore, as Figure 7 (a) and Figure 7 As shown in (b), longitudinal (the Y direction) stripes (composed of the third oxide layer 1102 and the second nitride layer 1104) for defining the access transistor and the element line, and a cross-point square for defining the active region will be formed, wherein the active region is the cross-point square located between the two longitudinal stripes, and Figure 7 (a) is along the Figure 7 (b) is a cross-sectional view taken along the cutting line in the X direction.
[0183] In step 122, the photoresist layer 1106 is retained to etch away the pad nitride layer 206, but the pad oxide layer 204 is retained, and then the photoresist layer 1106 and the pad oxide layer 204 are removed. Therefore, the horizontal silicon surface 208 (ie, HSS) is exposed as shown in FIG. Figure 7 (b) shows the cross-point square area corresponding to the active area.
[0184] In step 124, if Figure 8As shown, the horizontal silicon surface 208 exposed in the intersection square region is etched by the anisotropic etching technique to form a recess (e.g., the U-shaped recess), wherein the U-shaped recess is used to form a U-shaped channel 1302 of the access transistor, and the vertical depth of the U-shaped recess can reach about 60 nanometers, for example, starting from the horizontal silicon surface 208. Because the U-shaped recess of the access transistor is exposed, the U-shaped channel 1302 in the U-shaped recess can be doped with a properly designed concentration of boron (p-type dopant) to achieve channel doping, wherein the channel doping is to enable the access transistor to have a desired threshold voltage after a subsequent high-k metal-gate structure is formed. Then, a high-k dielectric insulating layer 1304 is formed at the bottom and sidewalls of the U-shaped recess, wherein the high-k dielectric insulating layer 1304 serves as a gate dielectric layer of the access transistor. A gate material 1306 is then selected that is suitable for the cell line conductivity and that can achieve the targeted work-function performance of the access transistor. Figure 8 As shown, the gate material 1306 must be thick enough to fill the U-shaped recess, and the gate material 1306 is etched back to form a longitudinal (the Y direction) element line.
[0185] In step 126, if Figure 8 As shown, a third nitride layer 1402 (i.e., a dielectric cap) is deposited, and then a fourth oxide layer 1404 is deposited, wherein the third nitride layer 1402 and the fourth oxide layer 1404 are stacked together to make their total thickness sufficient to fill the gap between the two adjacent longitudinal stripes (composed of the third oxide layer 1102 and the second nitride layer 1104) and finally form a flat surface. Then, the fourth oxide layer 1404 is etched back (or polished) to make the top of the fourth oxide layer 1404 flush with the top of the second nitride layer 1104, thereby forming a composite layer composed of the fourth oxide layer 1404 and the third nitride layer 1402 directly above the element line (i.e., the gate material 1306).
[0186] In step 128, if Figure 8As shown, the second nitride layer 1104 is etched away by the anisotropic etching technique, and the fourth oxide layer 1404 / third nitride layer 1402 is left on the top of the word line. Then the third oxide layer 1102 is also etched away by the anisotropic etching technique to expose the pad nitride layer 206. The gate structure (e.g., the fourth oxide layer 1404 / third nitride layer 1402 / gate material 1306) simultaneously implements the gate structure of the U-shaped transistor in the U-shaped recess and the word lines (i.e., the first word line word line-1, the second word line word line-2, and the third word line word line-3) in the longitudinal direction (i.e., the Y direction).
[0187] In step 130, if Fig. 9 As shown, the pad nitride layer 206 is removed to leave the pad oxide layer 204. The second oxide layer 1002 (ie, CVD-STI-oxide2) is etched back to make the top of the second oxide layer 1002 flush with the top of the pad oxide layer 204.
[0188] Then, in step 132, as Fig. 9 As shown, a fourth nitride layer 1602 is deposited and etched by the anisotropic etching technique to generate a fourth nitride spacer with a carefully designed thickness. A fifth oxide layer 1604 is then deposited and etched by the anisotropic etching technique to generate a fifth oxide spacer. A fifth nitride layer 1606 is then deposited. Fig. 9 The fifth nitride layer 1606 is then etched by the anisotropic etching technique to create a fifth nitride spacer so that the fifth nitride spacer surrounds the word line (eg, the first word line-1, the second word line-2, the third word line-3).
[0189] like Fig. 9As shown, for the convenience and clarity of describing the memory array circuit having the word lines and the bit lines, the word line located in the center is marked as the first word line word line-1 (corresponding to the access transistor AQ1), the word line adjacent to the left of the first word line word line-1 is marked as the second word line word line-2 (corresponding to the access transistor AQ2 adjacent to the left of the access transistor AQ1), and the pad oxide layer 204 still covers the area between the first word line word line-1 and the second word line word line-2 as the drain area to reserve for the first drain Drain-1 of the access transistor AQ1 and the second drain drain-2 of the access transistor AQ2. The word line adjacent to the right of the first word line word line-1 is marked as the third word line word line-3 (corresponding to the access transistor AQ3 adjacent to the right of the access transistor AQ1), and the pad oxide layer 204 still covers the area between the first word line word line-1 and the third word line word line-3 as the source region to reserve the first source source-1 of the access transistor AQ1 and the third source source-3 of the access transistor AQ3. Then, in step 134, Fig. 9 A sixth oxide layer 1802 is deposited over the entire surface as shown.
[0190] In step 136, if Fig.10 As shown, a spin-on dielectric 1902 is deposited, wherein the thickness of the spin-on dielectric 1902 is sufficient to fill the gaps between the above-mentioned word lines (the first word line word line-1, the second word line word line-2, and the third word line word line-3), and then the spin-on dielectric 1902 is polished by the chemical mechanical planarization technology to make the top of the spin-on dielectric 1902 flush with the top of the sixth oxide layer 1802. Then, a photoresist is deposited on the flush surface to cover the area reserved for the drain region (the first drain drain-1 and the second drain drain-2), and to expose the area reserved for the source region (the first source source-1 and the third source source-3). Then, as shown in FIG. Fig.10 As shown, the sixth oxide layer 1802 surrounding the above-mentioned word lines (the first word line word line-1, the second word line word line-2 and the third word line word line-3) can be used as a self-alignment mask to remove the spin-coated dielectric 1902 corresponding to the area reserved for the source region (the first source source-1 and the third source source-3).
[0191] Then, in step 138, Fig.10As shown, the exposed sixth oxide layer 1802 and the pad oxide layer 204 at the center of the source region between the two wordlines (the first wordline-1 and the third wordline-3) are etched away to expose the horizontal silicon surface 208. Because the exposed horizontal silicon surface 208 is located between the first source source-1 of the access transistor AQ1 and the third source source-3 of the access transistor AQ3, the horizontal silicon surface 208 between the first source source-1 and the third source source-3 is called the horizontal silicon surface HSS-1 / 3. Fig.10 As shown, the horizontal silicon surface HSS-1 / 2 between the first element line wordline-1 and the second element line wordline-2 is used as the region of the first drain drain-1 (that is, the drain of the access transistor AQ1) and the second drain drain-2 (that is, the drain of the access transistor AQ2), and is also used as the region vertically connecting the access transistors AQ1 and AQ2 to the subsurface bit line 902. In addition, to the right of the first element line wordline-1, the horizontal silicon surface HSS-1 / 3 between the first element line wordline-1 and the third element line wordline-3 is used as the region of the first source source-1 (that is, the source of the access transistor AQ1) and the third source source-3 (that is, the source of the access transistor AQ3), but because the first source source-1 and the third source source-3 will be connected to the memory cell storage nodes CSN1 and CSN3 (not shown) respectively. Fig.10 ), so the first source source-1 and the third source source-3 must be separated and cannot be connected.
[0192] In addition, the above-mentioned lithography technology is used to cover the horizontal silicon surface HSS-1 / 2, but the mask used in the lithography technology is not a critical mask, and its function is only to separate the process of processing the horizontal silicon surface HSS-1 / 3 from the process of processing the horizontal silicon surface HSS-1 / 2. As mentioned above, the thickness of the deposited spin-on dielectric 1902 must be sufficient to form a smooth surface morphology, and then a photoresist is deposited as a mask to protect the spin-on dielectric 1902 covering the drain region (the first drain drain-1 and the second drain drain-2) but expose the spin-on dielectric 1902 covering the source region (the first source source-1 and the third source source-3).
[0193] like Fig.10 As shown, the silicon material below the horizontal silicon surface HSS-1 / 3 (corresponding to the center of the source region) is excavated by the anisotropic etching to generate a hole hole-1 / 3 (e.g., 140 nanometers deep), wherein two opposite sides of the hole hole-1 / 3 (not shown in FIG. Fig.10 ) are respectively surrounded by the first nitride spacer layer and the first oxide spacer layer at the lower edge, and the other two pairs of sides of the hole hole-1 / 3 are surrounded by the p-type substrate 202.
[0194] Then, in step 140, as Fig.10 As shown, the seventh oxide layer 2102 is thermally generated to fill the hole hole-1 / 3, and the isotropic etching is used to accurately remove all the seventh oxide layer 2102 above the horizontal silicon surface 208, but the seventh oxide layer 2102 filling the hole hole-1 / 3 is left, wherein the seventh oxide layer 2102 filling the hole hole-1 / 3 is called a seventh oxide plug (oxide-7 plug), wherein the seventh oxide plug has a smooth surface and its height is flush with the top of the pad oxide layer 204.
[0195] Then, in step 142, as Fig.10 As shown, a spin-coated dielectric 2202 is deposited, wherein the thickness of the spin-coated dielectric 2202 is sufficient to fill the vacancy above the top of the seventh oxide layer 2102 in the hole 1 / 3, and then the top material of the spin-coated dielectric 2202 is removed by the chemical mechanical planarization technology so that the top of the spin-coated dielectric 2202 is the same height as the top of the fourth oxide layer 1404.
[0196] In step 144, if Fig.11 As shown, the photoresist 2302 is deposited by the lithography technology to cover the area corresponding to the source region and expose the area reserved for the drain region, wherein the mask used in step 144 is not a critical mask, and its function is only to separate the process of processing the horizontal silicon surface HSS-1 / 3 from the process of processing the horizontal silicon surface HSS-1 / 2. Then the exposed spin-on dielectric 1902, the exposed sixth oxide layer 1802, and the exposed liner oxide layer 204 are removed to expose the horizontal silicon surface (that is, the horizontal silicon surface HSS-1 / 2). Then, the silicon material corresponding to the horizontal silicon surface HSS-1 / 2 is excavated and removed by the anisotropic etching technology to generate a hole hole-1 / 2, wherein two pairs of sides of the hole hole-1 / 2 are respectively surrounded by the p-type substrate 202, and the third side of the hole hole-1 / 2 (not shown) is Fig.11 ) is surrounded by the first nitride spacer layer at the lower edge, and the fourth side of the hole hole-1 / 2 (not shown in FIG. Fig.11 ) is surrounded by the first oxide spacer layer, and the third side and the fourth side of the hole hole-1 / 2 are further surrounded by the second oxide layer 1002 (ie, CVD-STI-oxide2, not shown in FIG. Fig.11 ) at the external limit.
[0197] Then, in step 146, if Fig.11As shown, the photoresist is removed, and then an eighth oxide layer 2402 is thermally generated to produce an eighth oxide spacer, wherein among the four sidewalls of the hole hole-1 / 2, except for the third sidewall which is covered by the first nitride spacer at the lower edge, the eighth oxide spacer covers the remaining three sidewalls of the four sidewalls of the hole hole-1 / 2 and the bottom of the hole hole-1 / 2. In addition, Fig.12 is a schematic diagram illustrating a cross-sectional view of the memory array circuit along the Y2 direction, wherein the Y2 direction extends along the center of the hole Hole-1 / 2 and is perpendicular to the X direction.
[0198] Then, in step 148, if Fig.13 The lower edge first nitride spacer on the third sidewall in the hole hole-1 / 2 is removed by the isotropic etching technique (because the lower edge first nitride spacer is very thin, the isotropic etching technique will not damage other structures on the horizontal silicon surface 208, and will not remove the eighth oxide layer 2402 in the hole hole-1 / 2).
[0199] Then, in step 150, as Fig.13 As shown, an n+ heavily doped polysilicon film 2802 is deposited, wherein the thickness of the n+ heavily doped polysilicon film 2802 is sufficient to fill the hole hole-1 / 2, and then the n+ heavily doped polysilicon film 2802 on the horizontal silicon surface 208 is etched back by the isotropic etching technology to leave an n+ polysilicon plug (n+polysiliconplug) in the hole hole-1 / 2, and the n+ polysilicon plug is connected to the undersurface bit line (UGBL) through the opening on the third sidewall of the hole hole-1 / 2, wherein the third sidewall of the hole hole-1 / 2 was originally covered by the first nitride spacer at the lower edge.
[0200] like Fig.13 As shown, the connection between the n+ polysilicon plug and the UGBL is from the sidewall of the n+ polysilicon plug to the sidewall of the tungsten interconnect (that is, the UGBL) in the hole Hole-1 / 2, wherein the n+ polysilicon plug and the UGBL are both conductive materials and connected in a sidewall self-aligned manner, but are completely isolated from the p-type substrate 202 by the eighth oxide layer 2402. Such an n+ polysilicon plug is a vertical interconnection line, or an interconnection line extending upward or downward.
[0201] Then, in step 152, if Fig.14As shown, an appropriate amount (for example, a height of 20 nm below the horizontal silicon surface 208) of the upper half of the eighth oxide layer 2402 is removed by the anisotropic etching technology, resulting in the height of the eighth oxide spacer (eighth oxide layer 2402) being lower than the height of the n+ polysilicon plug.
[0202] Then, in step 154, if Fig.15 As shown, an n+ polysilicon layer 3202 is grown from the silicon seed provided by the n+ polysilicon plug by the selective epitaxial growth (SEG) technology, resulting in a necklace-type conductive n+ polysilicon layer 3202 (called an n+ collar) connected to the horizontal silicon surface 208 on both sides of the hole hole-1 / 2, which can serve as the first drain drain-1 of the access transistor AQ1 and the second drain drain-2 of the access transistor AQ2, respectively, and can also serve as a conductive bridge between the subsurface bit line (UGBL) and the access transistors AQ1 and AQ2.
[0203] Then, if Fig.15 As shown, a ninth oxide layer 3302 with a well-designed thickness is locally thermally generated on the n+ collar to cover the horizontal silicon surface HSS-1 / 2, but the n+ collar is retained to serve as the first drain drain-1 of the access transistor AQ1 and the second drain drain-2 of the access transistor AQ2, respectively. Such a conductive plug (including the n+ polysilicon plug and the n+ collar) is a vertical interconnection line connected to a horizontally extending subsurface interconnection (subsurface bit line (UGBL)).
[0204] Fig.16Schematic diagram of realizing the vertical connection (bridge) of the access transistor and the drain region and how they are connected to the subsurface interconnection line disclosed in another embodiment of the present invention. The eighth oxide layer 2402 is thermally generated by following the above process and the first nitride spacer layer at the lower edge in the hole hole-1 / 2 is removed (instead of depositing and etching back the n+ heavily doped polysilicon film 2802 to leave the n+ polysilicon plug in the hole hole-1 / 2), and the hole hole-1 / 2 is filled with tungsten or other metal material (by deposition or any other means), and then etched back to leave a tungsten plug (called W-2 plug) in the hole hole-1 / 2. The W-2 plug is connected to the subsurface interconnection through its opening located on the sidewall of the hole hole-1 / 2 covered by the first nitride spacer layer. The height of the W-2 plug is lower than the horizontal silicon surface 208 by a well-designed distance (for example, about 20nm). The exposed portion of the eighth oxide layer 2402 within the hole hole-1 / 2 but not covered by the W-2 plug is removed using the anisotropic etching technique. Then, an n+ polysilicon layer (e.g., n+ heavily doped polysilicon film 2802) is deposited to fill the hole hole-1 / 2, and then the n+ polysilicon layer is etched back to make the top of the n+ polysilicon layer flush with the horizontal silicon surface 208, and the necklace-shaped conductive n+ polysilicon layer connected to the horizontal silicon surface 208 on both sides of the hole hole-1 / 2 can serve as the first drain drain-1 of the access transistor AQ1 and the second drain drain-2 of the access transistor AQ2, respectively, and can also serve as a conductive bridge between the undersurface bit line (UGBL) 902 and the access transistors AQ1 and AQ2. Then, a ninth oxide layer 3302 is locally thermally grown over the tungsten plug to cover the horizontal silicon surface HSS-1 / 2 but leave the n+ collar as the first drain drain-1 of the access transistor AQ1 and the second drain drain-2 of the access transistor AQ2.
[0205] The following is an embodiment of the present invention to explain the formation of the drain region / source region of the access transistor. In step 45, as Fig.17As shown, the sixth oxide layer 1802 and the fifth nitride layer 1606 are removed, and then n-type doped drain regions and source regions are formed in the p-type substrate 202, respectively (for example, the n-type dopant injected through the pad oxide layer 204 can be used to create the drain region and the source region with an np junction (that is, the first drain Drain-1 and the first source Source-1 of the access transistor AQ1, the second drain Drain-2 of the access transistor AQ2, and the third source Source-3 of the access transistor AQ3) in the p-type substrate 202). In addition, a rapid thermal annealing (RTA) process step is required to activate the n-type dopant to eliminate any defects caused by ion implantation. Of course, other methods can also be used to form the drain region and the source region, such as selective generation / selective epitaxial generation. In addition, n-type lightly doped drains (NLDD) can be formed under the fifth oxide spacer and the fourth nitride spacer. Afterwards, a conventional stack capacitor or other type of capacitor (not shown) may be formed to connect the first source Source-1 of the access transistor AQ1 to complete the memory cell (ie, dynamic random access memory cell) in the memory array circuit.
[0206] Therefore, as described above, a subsurface bit line (or subsurface interconnect) is formed and connected to the drain region of the access transistor (e.g., access transistor AQ1). In addition, some drain regions or source regions of different access transistors (e.g., access transistors AQ3, AQ2) can be connected to the subsurface interconnect that provides the back power supply Vdd or Vss. In addition, in another embodiment of the present invention, according to the above process, two or more different subsurface interconnects can also be formed at two or more different vertical heights in the pipeline, and separated from each other by isolation materials (e.g., Fig.18 The different subsurface interconnects may have different lengths and may be connected to different transistors. In another embodiment of the present invention, the different subsurface interconnects in the pipeline may be connected to different transistors. Fig.18 In addition, the different subsurface interconnects may be made of different materials and may have different thicknesses.
[0207] Because the subsurface interconnect can have a similar resistance to the metal 0 used in the fab, as described above, the subsurface interconnect can be used as a bit line in the memory array circuit to free up layout space and reduce the size of the memory array circuit. Fig.19 , Fig.19is a schematic diagram illustrating a conventional dynamic random access memory structure, wherein write data XIO (e.g., signal "ONE" or a high potential signal) is transmitted along the data input circuit DI, the global input / output path GIO, the data line sense amplifier 70, and the data line DL. In addition, the write data XIO is transmitted between the data line DL and the memory array 75, wherein the write data XIO is stored in the corresponding storage node in the memory array 75 through the bit line BL. In the memory array 75, as shown in FIG. Fig. 20 As shown, the sense amplifier 80 is connected to the bit line BL, wherein the bit line BL is coupled to the data line DL through the bit switch BS100. A plurality of dynamic random access memory cells (e.g., 512 dynamic random access memory cells) are connected to the bit line BL. A first end E1 of the bit line BL is connected to a first dynamic random access memory cell among the plurality of dynamic random access memory cells, and a second end E2 of the bit line BL is connected to a last dynamic random access memory cell among the plurality of dynamic random access memory cells. Fig.21 In the embodiment, a dynamic random access memory cell including an access transistor 11 and a storage capacitor 12 is taken as an example. The gate of the access transistor 11 is coupled to the bit line WL, and the sense amplifier 80 is coupled to the access transistor 11 through the bit line BL. The sense amplifier 80 includes two P-type metal oxide semiconductor transistors ( Fig.21 P1 and P2 in the figure) and two N-type metal oxide semiconductor transistors ( Fig.21 N3 and N4 in FIG. 1 ). In addition, the sense amplifier 80 is connected to the VSSSA power supply through the transistor SLP, wherein the transistor SLP is connected to the P-type metal oxide semiconductor transistors P1 and P2 through the connection line LSLP. In addition, the sense amplifier 80 is also connected to the VSS power supply through the transistor SLN, wherein the transistor SLN is connected to the N-type metal oxide semiconductor transistors N1 and N2 through the connection line LSLN. The dynamic random access memory cell uses the access transistor 11 as a switch to control the charge stored from the bit line BL to the storage capacitor 12 in the write (WRITE) mode or the charge transferred from the storage capacitor 12 to the bit line BL in the read (READ) mode. In addition, as Fig.19 As shown, VCCSA is the supply voltage.
[0208] Thus, the conventional DRAM cell array design is as follows Fig.19 , Fig. 20 , Fig.21 As shown, Fig.21The invention comprises (1) a dynamic random access memory array having a plurality of dynamic random access memory cells, such as the most popular 1T1C cell, wherein the 1T1C cell comprises an access transistor (i.e., 1T, whose threshold voltage Vth is usually around 0.7V) and a storage capacitor (i.e., 1C, which is usually 17fF in a 20-28nm process node, and may be less than 10fF if the process node is lower than 20nm); (2) the drain regions of the access transistors of the plurality of 1T1C cells are respectively connected to interconnection lines (the interconnection lines are called bit lines); (3) the gates of the access transistors of the plurality of 1T1C cells are also respectively connected through interconnections called bit lines. The bit lines are connected to a sense amplifier (i.e., sense amplifier 80) of a cross-coupled circuit such as a complementary metal oxide semiconductor (CMOS). Accordingly, there is another bit line, called a complementary bit line (bit line-Bar), which carries a signal complementary to the signal of the bit line BL and is also connected to the same sense amplifier (i.e., sense amplifier 80). Along the interconnection between the bit line BL and the complementary bit line BLB, other devices are also connected to perform the complete bit line function in operation, such as an equalization circuit 2203 (e.g., an equalization circuit 2203 for equalizing voltage potentials) Fig. 22 as shown) and for controlling the bit lines ( Fig. 22 BL and BLB) and data lines ( Fig. 22 The bit switch ( Fig. 22 BS1 and BS2 shown in FIG. 1 ). In addition, Fig. 22 As shown in FIG. 2 , SLNS and eqpl are control signals for the sensing amplifier and the equalization circuit 2203, respectively. Fig. 20 As shown, VCCSA and VSS are supply voltages, and EN1 and EN2 are control signals of the sense amplifier 80. Fig. 22 As shown, VSSSA and VBL are voltages.
[0209] Fig. 2222 is a schematic diagram of a conventional sense amplifier design, wherein the sense amplifier has connection lines 1 to 23 for circuit connection. Bit switches BS1 and BS2 are connected to bit line BL and complementary bit line BLB through connection lines 11 and 12, respectively. P-type metal oxide semiconductor transistor P1 is connected to bit line BL and complementary bit line BLB through connection lines 1 and 6, respectively. P-type metal oxide semiconductor transistor P2 is connected to bit line BL and complementary bit line BLB through connection lines 5 and 2, respectively. N-type metal oxide semiconductor transistor N3 is connected to bit line BL and complementary bit line BLB through connection lines 3 and 7, respectively. N-type metal oxide semiconductor transistor N4 is connected to bit line BL and complementary bit line BLB through connection lines 8 and 4, respectively. The three transistors in equalization circuit 2203 are connected to bit line BL and complementary bit line BLB through connection lines 9, 10, 13, and 14.
[0210] On the other hand, Fig.23 , Fig.24 As shown, in a conventional DRAM open-bitline's bit line sense amplifier (BLSA) design, two bit line sense amplifiers (eg Fig.23 , Fig.24 The bit line sense amplifiers BLSA0, BLSA1 in the bit line sense amplifier BLSA0 are placed side by side. The connection lines 1 to 23 in the bit line sense amplifier BLSA0 and the connection lines 1a to 23a in the bit line sense amplifier BLSA1 usually use metal 0 (metal zero layer) or other metal layers during the layout process, and the bit line BL and the complementary bit line BLB also use metal 0. Therefore, those connection lines (connection lines 1 to 23 and connection lines 1a to 23a) made of metal 0 will incur spacing restrictions therebetween and produce a larger layout area.
[0211] However, if Fig.25 As shown, the present invention proposes a subsurface bit line technology, and then connects the subsurface bit line to the above-mentioned bit line sense amplifier. In the present invention, Fig.24 The connection lines 1 to 23 and Fig.24 The connection lines 1a-23a in the bit line sense amplifier BLSA1 are still made of metal 0 or other metal layers like a conventional dynamic random access memory, but the subsurface bit lines (such as the bit line BL and the complementary bit line BLB) are now located below the horizontal silicon surface 208. Therefore, the connection lines made of metal 0 in the bit line sense amplifier BLSA can be connected to the subsurface bit lines through the switch layer. Fig.25As shown, the subsurface bit line is connected to the access transistor of the active area AA (e.g., N+AA) through the connecting pin therein. In one embodiment of the present invention, the active area AA is a silicon area surrounded by the isolation area. Therefore, during the layout process, the subsurface bit line (e.g., the bit line BL and the complementary bit line BLB) can be connected to the active area AA first, and then, as shown in FIG. Fig.25 As shown at the bottom of FIG. , the active area AA is connected to metal 0 through metal 0 contact during the layout process. Fig.18 As shown, the two subsurface bit lines UGBL may be located in one isolation region and / or in two different layers in the isolation region. Fig.23 and Fig.24 shown) can be relaxed.
[0212] In one embodiment of the present invention, the switch layer may include a metal 0 contact, an active area AA connected to the metal 0 contact, and a connection plug connecting the active area AA to the subsurface bit line. Through the connection plug, the active area AA and the metal 0 contact, the connection line made of metal 0 in the bit line sense amplifier BLSA can be electrically connected to the subsurface bit line (e.g., the bit line BL and the complementary bit line BLB). In addition, the active area AA in the switch layer can be an active area without any transistors therein. As described above, the connection plug in the active area AA (see Fig.15 In addition, in order to reduce the resistance of the switch layer, two or more connecting plugs in the active area AA can be connected to the same subsurface bit line (such as Fig.25 ).
[0213] In addition, if Fig.26 As shown in FIG. 1 , in a conventional bit line sense amplifier, the height of a bit line sense amplifier is 4 times the bit line pitch. When the technology node λ is between 20 and 30 nanometers (nm) (e.g., 25nm), the bit line pitch is about 80nm (when λ=25nm, 80nm is about 3.2λ), so the height of the bit line sense amplifier is about 80nmx4=320nm or 0.32um. In addition, Fig.24As shown, the connection lines 1 to 14 in the bit line sense amplifier BLSA0 and the connection lines 1a to 14a in the bit line sense amplifier BLSA1 are usually made of metal 0 layer during the layout process, so there will be spacing limitations between the connection lines 1 to 14 made of metal 0 and between the connection lines 1a to 14a, resulting in a larger layout area. Moreover, for the bit line sense amplifier BLSA0, the length of the complementary bit line BL0B is greater than the length of the bit line BL0. Of course, for the bit line sense amplifier BLSA1, the length of the bit line BL1 is greater than the length of the complementary bit line BL1B. Therefore, the different lengths between the bit line / complementary bit line may cause the input of the bit line sense amplifier to be inconsistent. Worse, the capacitive coupling between the bit line BL1 and the complementary bit line BL0B also increases the bit line capacitance.
[0214] In addition, in another embodiment of the present invention, the present invention discloses using the above-mentioned underground interconnection line (UGIL) technology in the bit line sense amplifier (for example, most of the connection lines in the bit line sense amplifier can be made of underground interconnection lines instead of metal 0) and re-layout the bit line sense amplifier. In this way, not only the area of the bit line sense amplifier can be reduced, but also the bit line mismatch issue is solved and the bit line capacitance is reduced. For example, Fig. 27As shown, the connection lines 1 to 4, 9 to 23 in the conventional bit line sense amplifier can be used by an underground connector (UGC, or a vertical interconnection line or a connection pin as described above) or other contacts, so these underground connectors (UGC) are arranged below the horizontal silicon surface 208 and only four connection lines 5, 6, 7, and 8 are made of metal 0. That is to say, the underground interconnection technology will release the metal 0 space for other circuit layouts. Therefore, about 82% of the metal 0 in the conventional bit line sense amplifier is released (that is, the metal 0 of 19 connection lines (connection lines 1 to 4, 9 to 23) among the connection lines 1 to 23 is replaced by the underground interconnection line, so 19 / 23, about 82% of the metal 0 is released) for other circuit layouts. Since 82% of the wiring area in the metal 0 can be released (all replaced by the underground interconnection line), the wiring of the metal 0 will no longer be a problem of spacing limitation in the layout of the bit line sense amplifier. Therefore, the area of the bit line sense amplifier can be significantly reduced by reducing the size of the transistors in the bit line sense amplifier. In addition, since the connecting wires 1, 2, 3, 4, 9, 10, 11, 12, 13, 14 are connected to the bit line BL or the complementary bit line BLB, the bit line BL or the complementary bit line BLB can also be replaced by a subsurface interconnection line without the above-mentioned switching layer connecting the bit line BL and the complementary bit line BLB to the connecting wires 1, 2, 3, 4, 9, 10, 11, 12, 13, 14. Therefore, as Fig.28A As shown in FIG. 1 , according to the present invention, the layout area (0.32um height x 6um width) of a conventional bitline sense amplifier with a 4-bitline pitch can be reduced to the layout area (0.16um height x 6.5um width) of a bitline sense amplifier with a 2-bitline pitch, that is, the layout area is reduced by 46%. Fig.28B As shown, the two conventional bit line sense amplifiers placed side by side can be replaced by the two bit line sense amplifiers stacked vertically (i.e., stacked in the y direction) disclosed in the present invention. Therefore, it is worth noting that the obvious advantages of the application of the subsurface interconnect technology in the bit line sense amplifier are that the layout area of the bit line sense amplifier is reduced by 46%, the extra bit line length from the complementary bit line BLB0 to the bit line sense amplifier BLSA0 and the bit line BL1 to the bit line sense amplifier BLSA1 is eliminated (i.e., a more balanced layout), the bit line capacitance and sensing power are reduced, and the sensing speed is improved.
[0215] Fig.291 is a schematic diagram illustrating a layout result of a bit line sense amplifier disclosed in the present invention (1-bit bit line sense amplifier BLSA0_new), wherein the bit line BL and the complementary bit line BLB are the underground interconnection line UGIL2, and the connecting lines 1, 2, 3, 4, 9, 10, 11, 12, 13, 14 are vertical interconnection lines or underground connectors (Underground Connector, UGC) connected to the bit line BL and the complementary bit line BLB in parallel, wherein the bit line BL and the complementary bit line BLB are horizontal underground interconnection lines. Fig.29 As shown, Fig.29 The top part of the figure is the layout result, where BS represents the bit switches on both sides and PSA represents Fig. 22 The P1 and P2 transistors shown, NSA stands for Fig.21 The N3 and N4 transistors are shown. Fig. 22 In the above table, SLN stands for Fig. 22 SLN transistor in, LSLN / LSLP means Fig.21 and Fig. 22 The LSLN / LSLP lines in the EQBL representation Fig. 22 The equalization circuit 2203 in.
[0216] In addition, if Fig.29 As shown, the connection lines 15 and 16 (or local data lines in the bit line sense amplifier) respectively connected to the external (or global) data lines DL and DLB in the present invention are also replaced by another horizontal subsurface interconnection line UGIL1 (UGIL_Layer1), and the same is true for the connection line LSLN. Therefore, the connection lines 15 and 16 (or local data lines LDL / LDLB in the bit line sense amplifier) are connected to the bit switch BS through vertical interconnection lines or subsurface connectors (since these subsurface connectors are connected to the subsurface interconnection line UGIL1, these subsurface connectors are Fig.29 denoted as UGC1 in FIG. 1 ), and connecting lines 17 to 22 are also vertical interconnects or subsurface connectors UGC1 that are electrically connected to connecting LSLN / LSLP and connecting VBL. Similarly, in Fig.29 The subsurface connector connected to the subsurface interconnect UGIL2 is denoted as UGC2. Fig.29 As shown in the top view layout result, the transistor of the bit line sense amplifier is set between the horizontal spacing between the bit line (UGIL2) and the local data line (UGIL1), the group of bit line switches is set between the horizontal spacing between the bit line (UGIL2) and the local data line (UGIL1), and the transistor of the equalization circuit is also set between the horizontal spacing between the bit line (UGIL2) and the local data line (UGIL1).
[0217] like Fig.29As shown, UGIL2 (UGIL_Layer2) and UGIL1 can be subsurface interconnects of different vertical heights and relatively displaced in the X and / or Y directions. For example, UGIL1 is higher (or closer to the horizontal silicon surface (HSS)) than UGIL2. At this time, UGIL2 and UGIL1 may not pass through metal 0, but may be connected to other higher metal layers (e.g., metal 1, metal 2, etc.) through elevated conductor pillars (ECoP). For example, ECoP2 is used to connect UGIL2 to other higher metal layers without passing through metal 0, and ECoP1 is used to connect UGIL1 to other higher metal layers without passing through metal 0. Therefore, as Fig.29 As shown, connection lines 15 and 16 (UGIL1) are connected to external data lines DL and DLB (metal 1) through ECoP1. How to form ECoP2 or ECoP1 from a lower metal layer (such as UGIL2 or UGIL1) to other higher metal layers (such as metal 1 or metal 2) can refer to U.S. Patent Application No. 17528481 (Name: MANUFACTURE METHOD FOR INTERCONNECTION STRUCTURE, Date: November 17, 2021), which is cited in its entirety.
[0218] therefore, Fig. 30A is a schematic diagram illustrating the layout result of two bit line sense amplifiers (BLSA0_new and BLSA1_new) for 2 bits stacked vertically, and Fig. 30B is a schematic diagram illustrating the layout result of four bit line sense amplifiers for 4 bits stacked vertically. It is worth noting that Fig. 30A , UGIL2 of the bit line sense amplifier BLSA1_new is below UGIL1 of the bit line sense amplifier BLSA0_new. In addition, Fig. 30B The step-by-step partial layout results can be found in Fig.33A , Fig.33B , Fig.33C , Fig.33D and Fig.33E In addition, it is worth noting that Fig. 30B It has the following characteristics:
[0219] (1) All bit line sense amplifiers are connected using pins;
[0220] (2) The device sequence along the subsurface interconnect line UGIL1 (UGIL_Layer1) and the subsurface interconnect line UGIL2 (UGIL_Layer2) can be placed in different sequences according to the top metal connection and circuit design requirements;
[0221] (3) 2 layers of subsurface interconnects for connecting pins 1 and 2;
[0222] (4) The buried oxide layer under the transistor source / drain reduces the channel length of the device;
[0223] (5) Elevated conductor post contact (ECoP contact) can be used for data line connection of each bit line sense amplifier bit, which is directly connected to any metal layer to reduce connection area / pitch;
[0224] (6) Each elevated conductor post of some voltage source contacts a common 4 bit line sense amplifier bits.
[0225] In addition, if Fig.31 and Fig.32 As shown, the transistors in the bit line sense amplifier (BLSA0_new and BLSA1_new) disclosed in the present invention may be transistors having a buried oxide layer (hereinafter referred to as BO (buried oxide) transistor) under the source / drain of the transistor, and the gate, drain and source of the BO transistor are smaller than the gate, drain and source of the transistor in the conventional bit line sense amplifier. For example, Fig.31 As shown in (a), the traditional contact-connected metal-oxide-semiconductor (MOS) transistor is connected to metal 0 through its contact, and its minimum height (400nm) is related to its gate length (90nm), contact size (50nm), and active area contact rules (24nm, 40nm, 82nm). In addition, as Fig.31 As shown in (b), a metal oxide semiconductor transistor with a subsurface connector (UGC) connection is connected to its subsurface connector through a subsurface interconnect line, wherein its minimum height (120nm) is related to its gate length (45nm) and the rule of the active region with the subsurface connector, and its minimum height (120nm) is less than Fig.31 (a) shows the minimum height (400 nm) of a conventional contact-connected metal oxide semiconductor transistor. For information on how to form a BO transistor, please refer to U.S. Patent Application No. 17,318,097 (title: COMPLEMENTARY MOSFET STRUCTURE WITH LOCALIZED ISOLATIONS IN SILICON SUBSTRATE TO REDUCE LEAKAGES AND PREVENT LATCH-UP, date: May 12, 2021), which is hereby cited in its entirety.
[0226] In summary, by using the subsurface interconnect technology disclosed in the present invention, the layout area (0.32um height x 6um width) of the conventional bitline sense amplifier with a 4-bitline pitch can be reduced to the layout area (0.16um height x 6.5um width) of the bitline sense amplifier with a 2-bitline pitch, that is, the layout area is reduced by 46%. Therefore, the two conventional bitline sense amplifiers placed side by side can be replaced by the two vertically stacked bitline sense amplifiers disclosed in the present invention. When the technology node λ is lower than 30nm (e.g., 25nm), the height of the bitline sense amplifier can be reduced from 3.2λ to 1.6λ. In addition, since all the bitline sense amplifiers are stacked vertically (i.e., stacked in the y direction) instead of placing two bitline sense amplifiers side by side, the bitline sense amplifier disclosed in the present invention does not have the problem of bitline mismatch when the technology node λ is lower than 30nm. In addition, compared with the conventional bitline sense amplifier, the bitline capacitance of the bitline sense amplifier disclosed in the present invention can also be reduced.
[0227] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A memory array circuit, characterized in that Include: A semiconductor substrate having an original surface; a bit line and a complementary bit line; and a bit line sense amplifier connected to the bit line and the complementary bit line, the bit line sense amplifier comprising: A first plurality of transistors, each transistor comprising a gate node, a first conductive node, and a second conductive node; and a first set of connection lines connecting first conduction nodes of the first plurality of transistors to the bit line and the complementary bit line; The first group of connection lines is located below the original surface of the semiconductor substrate.
2. The memory array circuit as claimed in claim 1, characterized in that The first set of connection lines is located below the original surface of the semiconductor substrate, and at least a portion of the first set of connection lines is isolated from the semiconductor substrate.
3. The memory array circuit as claimed in claim 1, characterized in that The bit line and the complementary bit line are located below an original surface of the semiconductor substrate and isolated from the semiconductor substrate.
4. The memory array circuit of claim 3, further comprising a second set of connection lines connecting the gate nodes of the first plurality of transistors to the bit lines and the complementary bit lines, wherein the second set of connection lines is located above the original surface of the semiconductor substrate.
5. The memory array circuit as claimed in claim 4, characterized in that Each connection line in the first set of connection lines extends from the original surface downward to the bit line or the complementary bit line, and the bit line or the complementary bit line extends laterally.
6. The memory array circuit as claimed in claim 1, characterized in that An equalization circuit is also included, wherein a third set of connection lines connecting the equalization circuit to the bit line and the complementary bit line is located below the original surface of the semiconductor substrate, and each connection line in the third set of connection lines extends downward from the original surface to the bit line or the complementary bit line.
7. The memory array circuit as claimed in claim 1, characterized in that The bit line sense amplifier further includes a local data line connected to a bit line switch and a local complementary data line connected to a complementary bit line switch, wherein the local data line and the local complementary data line lines are located below the original surface of the semiconductor substrate and isolated from the semiconductor substrate.
8. The memory array circuit as claimed in claim 7, characterized in that A fourth set of connection lines connecting the bit line switch and the complementary bit line switch to the bit line and the complementary bit line is located below the original surface of the semiconductor substrate, and each connection line in the fourth set of connection lines extends from the original surface down to the bit line or the complementary bit line.
9. The memory array circuit as claimed in claim 7, characterized in that The bit line is located below an original surface of the semiconductor substrate, and a distance between the original surface of the semiconductor substrate and the local data line is smaller than a distance between the original surface of the semiconductor substrate and the bit line.
10. The memory array circuit as claimed in claim 7, characterized in that The first plurality of transistors include two P-type metal oxide semiconductor transistors and two N-type metal oxide semiconductor transistors, a fifth group of connection lines connecting the second conductive nodes of the two P-type metal oxide semiconductor transistors is located below the original surface of the semiconductor substrate, and a sixth group of connection lines connecting the second conductive nodes of the two N-type metal oxide semiconductor transistors is located below the original surface of the semiconductor substrate.
11. The integrated circuit according to claim 10, characterized in that The fifth group of connection lines and the sixth group of connection lines extend laterally.
12. A memory array circuit, characterized in that Include: A semiconductor substrate having an original surface; a bit line and a complementary bit line, wherein the bit line and the complementary bit line are located below an original surface of the semiconductor substrate; and a bit line sense amplifier connected to the bit line and the complementary bit line, the bit line sense amplifier comprising: a first plurality of transistors coupled to the bit lines and the complementary bit lines; a set of bit line switches coupled to the bit line and the complementary bit line; a local data line and a local complementary data line coupled to the set of bit line switches; and an external data line and an external complementary data line coupled to the set of bit line switches through the local data line and the local complementary data line; The local data line and the local complementary data line are located below an original surface of the semiconductor substrate.
13. The memory array circuit as claimed in claim 12, characterized in that The first plurality of transistors and the set of bit line switches are disposed between a horizontal spacing between the bit line and the local data line.
14. The memory array circuit as claimed in claim 13, characterized in that Also included is an equalization circuit having a second plurality of transistors disposed between the horizontal spacing between the bit lines and the local data lines.
15. The memory array circuit as claimed in claim 14, characterized in that The horizontal spacing between the bit line and the local data line is less than 3.5λ, and λ is the technology node of the memory array circuit.
16. The memory array circuit as claimed in claim 15, characterized in that The technology node λ is below 30 nanometers.
17. The memory array circuit as claimed in claim 12, characterized in that The length of the bit line is substantially the same as the length of the complementary bit line.
18. The memory array circuit as claimed in claim 12, characterized in that A first set of connection lines connecting the first plurality of transistors to the bit lines and the complementary bit lines is located below the original surface of the semiconductor substrate, and a second set of connection lines connecting the first plurality of transistors to the bit lines and the complementary bit lines is located above the original surface of the semiconductor substrate.
19. The memory array circuit as claimed in claim 12, characterized in that A distance between an original surface of the semiconductor substrate and the local data line is shorter than a distance between the original surface of the semiconductor substrate and the bit line.
Citation Information
Patent Citations
Complementary mosfet structure with localized isolations in silicon substrate to reduce leakages and prevent latch-up
US20220320087A1
