Non-volatile memory device having a nitride-oxide dielectric layer

[0001] This application relates to U.S. patent application Ser. No. 11/197,668, filed on Aug. 4, 2005, entitled “Non-Volatile Memory Semiconductor Device Having an Oxide-Nitride-Oxide (ONO) Top Dielectric Layer.”

[0002] The present invention relates to a non-volatile memory device and a method for manufacturing a non-volatile memory device and, more particularly, to a non-volatile memory semiconductor device having a nitride-oxide dielectric layer and a method for manufacturing a non-volatile memory device having a nitride top dielectric layer.

[0003] Non-volatile memory (“NVM”) is a semiconductor memory device that is capable of storing the programmed data or information even when electricity supplied to it is removed from a device containing the NVM memory. Examples of NVM include Mask Read-Only Memory (Mask ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), and Electrically Erasable Programmable Read-Only Memory (EEPROM). Typically, NVM can be programmed with data, read and/or erased, and the programmed data can be stored for an extended period of time before being erased, such as ten years.

[0004] An example of an NVM includes Nitride Read Only Memory, which is a type of EEPROM that uses a silicon nitride layer as a charge-trapping layer to store data. Other charge-trapping materials may be used for forming NVM devices. An example of Nitride Read Only Memory cell design includes a metal-oxide-silicon field effect transistor (MOSFET) having an ONO (oxide-nitride-oxide) layer disposed between the gate and the source/drain of the semiconductor material. The nitride layer in the ONO layer is able to “trap” charge (electrons) when the device is “programmed.” Charge localization is the ability of the nitride material to store the charge without significant lateral movement of the charge throughout the nitride layer. Nitride Read Only Memory utilizes a relatively thick tunnel oxide layer, which typically negatively impacts the time it takes to erase a memory cell. Nitride Read Only Memory can be contrasted with conventional “floating gate” memory cells wherein the floating gate is conductive and the charge is spread laterally throughout the entire floating gate and charge is transferred through a tunnel oxide layer. Also, Nitride Read Only Memory can be repeatedly programmed, read, erased and/or reprogrammed by known voltage application techniques.

[0005] To provide a non-volatile memory device, it is therefore desirable to provide a non-volatile memory semiconductor device that has good charge retention ability or reduced charge loss compared to some of the conventional NVM, such as certain Nitride Read Only Memory devices. It is also desirable to provide an NVM that can be operated with adequate speed, such as that needed when erasing old information or in programming new information.

[0006] One example of the present invention comprises a non-volatile memory cell. The non-volatile memory cell includes a semiconductor substrate; a source region in a portion of the substrate; a drain region within a portion of the substrate, the drain region being spaced apart from the source region; a well region within a portion of the substrate, the well region surrounding at least a portion of the source and drain regions; a first carrier tunneling layer over the substrate; a charge storage layer over the electron tunneling layer; a second carrier tunneling layer over the charge storage layer; and a conductive control gate over the second carrier tunneling layer.

[0007] Another example of the invention may include a non-volatile memory cell including: a semiconductor substrate; a source region formed in a portion of the substrate; a drain region formed in a portion of the substrate and adjacent to the source region; a well region formed in a portion of the substrate and adjacent to the source region and drain region; an electron tunneling layer formed over the substrate; a charge storage layer formed over the electron tunneling layer; and a p-doped polysilicon layer over the charge storage layer.

[0008] Another example of the present invention may include a method of forming a non-volatile memory cell on a semiconductor substrate. The method includes: forming a source region in a portion of the substrate; forming a drain region within a portion of the substrate, the drain region being spaced apart from the source region; forming a well region within a portion of the substrate, the well region surrounding at least a portion of the source and drain regions; forming a first carrier tunneling layer over the substrate; forming a charge storage layer over the first carrier tunneling layer; forming a second carrier tunneling layer over the charge storage layer; and forming a conductive control gate over the second carrier tunneling layer.

[0009] The following drawings illustrate examples for illustrating the invention. However, it is noted that the invention is not limited to the precise arrangements, instrumentalities, scales, and dimensions shown in these examples, which are provided mainly for illustration purposes only. In the drawings:

[0010] FIG. 1 illustrates an example of a non-volatile memory cell having a single-layer hole-tunneling layer consistent the invention;

[0011] FIG. 2 illustrates an example of a non-volatile memory cell having a dual-layer hole-tunneling layer consistent the invention;

[0012] FIG. 3 illustrates an example of a non-volatile memory cell having a tri-layer hole-tunneling layer consistent the invention;

[0013] FIG. 4 illustrates an example of signal designations for separate terminals of a non-volatile memory consistent the invention;

[0014] FIGS. 5 and 6 illustrate one example of a programming operation of a non-volatile memory consistent the invention;

[0015] FIG. 7 illustrates one example of an erasing operation of a non-volatile memory consistent the invention;

[0016] FIG. 8A is a band energy diagram of a non-volatile memory consistent with the invention when the electric field is low; and

[0017] FIG. 8B is a band energy diagram of a non-volatile memory consistent with the invention when the electric field is high.

[0018] The following examples of NVM devices provide improved charge retention ability or reduced charge loss, which will now be described in detail. Referring to FIG. 1, an example of a non-volatile memory (NVM) cell 100 includes semiconductor substrate 102, source region 104 in a portion of substrate 102, and drain region 106 in a portion of substrate 102. Drain region 106 is spaced apart from source region 104. NVM cell 100 also includes well region 108 in a portion of substrate 102, and well region 108 surrounds at least a portion of source region 104 and drain region 106. For example, well region 108 may encompass at least the region between source region 104 and drain region 106 and the region below those two regions.

[0019] Additionally, NVM cell 100 includes electron, or first carrier, tunneling layer 110 formed over substrate 102 and covering at least a portion of the region between source region 104 and drain region 106. A charge storage layer 112 is formed over the electron tunneling layer 110 and hole, or second carrier, tunneling layer 114 is formed over the charge storage layer 112. A conductive control gate 116 is also formed over the hole tunneling layer 114. In the example of FIG. 1, NVM cell 100 is an N-channel device, which has both source region 104 and drain region 106 as n-type or n+-type regions surrounded by well region 108, a p-type region. Electron tunneling layer 110 provides electron tunneling during a programming operation and is directly above at least a portion of the region between source and drain regions 104 and 106. Hole tunneling layer 114 includes at least one dielectric layer and provides hole tunneling during an erasing operation. In one example, a memory device may be provided without hole tunneling layer 114 and with p-doped polysilicon, i.e. polysilicon with p-type dopants, as conductive control gate 116.

[0020] Thus, a p-type substrate, such as a p-type silicon substrate, may be used as substrate 102 and may itself provide p-type well region 108. It may also be doped or implanted with an n-type material in the areas where the desired source and drain regions are to be provided. Other techniques, such as trench formation techniques, may be used to provide any of the well, source, and drain regions 108, 104, and 106, and well region 108 may also be provided by a doping/implantation process to provide a p-type well in a non p-type substrate. Furthermore, NVM cell 100 may be modified into a P-channel device by using an n-type well with p-type source and drain regions.

[0021] Referring to FIG. 1, electron tunneling layer 110 may be a dielectric layer that allows electron tunneling at least during the programming operation of NVM cell 100. For example, electron tunneling layer 120 may be a silicon oxide layer, which may be between about 30-90Angstroms (Å) in thickness. Other than silicon oxide, aluminum oxide (Al₂O₃) may also be used. In some examples, a silicon oxide layer thicker than 30 Å may be used to prevent direct-tunneling charge loss from charge storage layer 112 after NVM cell 100 has been programmed (i.e., charge has been intentionally stored in charge storage layer 112). In other words, electron tunneling layer 110 also functions as an insulating layer for blocking charge loss during the charge-retention/non-programming phase of NVM cell 100.

[0022] Referring again to FIG. 1, charge storage layer 112 may be a silicon nitride layer in one example, which may be between about 50-150 Å in thickness. Other than silicon nitride (Si₃N₄), silicon oxynitride (SiO_xN_x), aluminum oxide (Al₂O₃), hafnium oxide (HfO₂), and the like may also be charge storage layer 112. In other words, any material providing appropriate charge-trapping characteristics may be charge storage layer 112 or a portion of it. In this example, charge storage layer 112 provides the programmable memory storage portion of NVM cell 100. Therefore, charge storage layer 112 may be a material that efficiently traps or stores charge after a programming operation, which may entail applying a programming voltage potential to control gate 125 and one of source region 104 and drain region 106 in order to induce or introduce charge into charge storage layer 112.

[0023] Referring to FIG. 1, hole tunneling layer 114 may contain one or more dielectric layers for providing hole tunneling at least during an erasing operation, which may include hole tunneling from conductive control gate 116 to charge storage layer 112. In the example of FIG. 1, hole tunneling layer 114 contains a single silicon oxide layer. The silicon oxide layer may be of about 15˜25 Å in thickness in one example. FIG. 2 illustrates a second example of hole tunneling layer 114, which contains silicon nitride layer 114b stacked over silicon oxide layer 114a. FIG. 3 illustrates a third example of hole tunneling layer 114, which contains a stacked structure of silicon oxide layer 114c, silicon nitride layer 114b, and silicon oxide layer 114a. These examples and other examples of hole tunneling layer, such as a band-gap-engineered tunneling dielectric layer, may provide an efficient tunneling barrier for hole injection from gate 116, yet may prevent or reduce tunneling leakage, which may involve electron loss, during desired charge retention, such as during a reading operation, a programming operation, or both. In some examples, each of the dielectric layers of hole tunneling layer 114 may be between about 10-30 Å in thickness. As an example, a tri-layer structure such as the one shown in FIG. 3 may have silicon oxide layer 114c of about 15˜25 Å in thickness, silicon nitride layer 114b of about 30˜60 Å in thickness, and silicon oxide layer 114a of about 15˜25 Å in thickness.

[0024] In some examples, hole tunneling layer 114 may contain at least silicon nitride or other insulating or dielectric materials with a low hole-barrier height, such as one of less than about 3 electron-volts (eV). As an example, silicon nitride is a relatively low hole barrier, which is on the order of 1.9 eV, so it may become “transparent” to hole tunneling under a relatively high electric field, such as when voltage applied to control gate 116 is greater than zero. The design of hole tunneling layer 114 may have it function as a tunneling dielectric for hole tunneling from control gate 116 during an erasing operation. A thin hole tunneling layer 114 may make the charge-trapping in the layer itself negligible during memory programming and charge-retention operations.

[0025] Conductive control gate 116 may be formed of undoped polysilicon, doped polysilicon, or a metal. For example, control gate 116 may be lightly or heavily n-doped polysilicon, lightly or heavily p-doped polysilicon, platinum (Pt), Tantalum Nitride (TaN) and the like. In some examples, a p-doped polysilicon may facilitates hole tunneling and provide good erasing characteristics. As illustrated above in one example, a memory device may be provided without hole tunneling layer 114 and with p-doped polysilicon as conductive control gate 116.

[0026] The paragraphs below illustrate an example of operating NVM cell 100 using an exemplary structure of NVM cell 100, such as the one illustrated in FIG. 2. The operation of NVM cell 100 may include programming information in to the cell, reading information already stored in the cell, and erasing information stored in the cell before writing or programming new information. Those operations may be achieved by applying appropriate voltages to separate terminals of NVM cell 100. FIG. 4 illustrates an example of signal designations. Referring to FIG. 4, gate voltage V_g is applied to control gate 116, source voltage V_s is applied to source 104, drain voltage V_d is applied to drain 106, and substrate voltage V_sub is applied to substrate 102.

[0027] FIG. 5 and 6 illustrate one example of a programming operation. In the illustrated example, NVM cell 100 is programmed by an electron injection method, such as a channel hot electron (CHE) injection. However, more than one programming methods may also be used. For n-channel devices, channel hot electron (CHE) injection, source side injection (SSI), channel initiated secondary electron (CHISEL) injection, and the like, can be employed. For p-channel device, band-to-band tunneling hot electron (BBHE) injection, channel hot-hole initiated hot electron injection, negative Fowler-Nordheim (−FN) gate injected electrons, or the like can be employed. By having at least a pair of source and drain regions, NVM cell 100 can be programmed with 2-bits per cell with a first (left) bit in charge storage layer 112 closer to source region 104 and a second (right) bit in charge storage layer 112 closer to drain region 106.

[0028] Referring to FIG. 5, a right bit or drain-side bit programming operation using CHE injection may entail applying a gate programming voltage V_g, such as a voltage of +10 volts, to control gate 116, a programming source voltage V_s of zero, and a programming drain voltage V_d of +5 volts. The substrate voltage V_sub may be zero. Referring to FIG. 6, similarly, a left bit or source-side bit programming operation using CHE injection may entail applying a gate programming voltage V_g, such as a voltage of +10 volts, to control gate 116, a programming source voltage V_s of +5 volts, and a programming drain voltage V_d of zero. The substrate voltage V_sub may be zero. FIGS. 5 and 6 are merely illustrative of one possible example, and different voltage combination may be used to place electrons in charge storage layer 112 in order to store information therein.

[0029] NVM cell 100 may be read by several reading methods. For example, it can be read by applying a gate read voltage to control gate 112 and a source/drain read voltage to one of source region and drain regions 104 and 106 and grounding the other one. The substrate voltage V_sub may be zero. As an example, the gate read voltage and the source/drain read voltage may each be between about 1-2 volts.

[0030] FIG. 7 illustrates one example of an erasing operation. Referring to FIG. 7, an erasing operation may entail applying a gate erasing voltage V_g, such as a voltage of +14 volts, to control gate 116 and a source/drain erasing voltage V_s/d, such as a voltage of 0 volts, to one or both of source region 104 and drain region 106. The substrate voltage V_sub may be zero. A voltage difference such as this exemplary one may induce charge into charge storage layer 112. As an example, a positive direct current voltage at control gate 116 may provide hole tunneling from control gate 116 toward charge storage layer 112. In some examples, a gate erasing voltage V_g is about +10 to about +18 volts. Additionally, a zero or negative source/drain erasing voltage may be applied to one or both of source region 104 and drain region 106. As an example, using hole-tunneling erase from the gate avoid the need to utilize band-to-band hot hole erase (BTBHHE), thereby reducing or eliminating hot-hole introduced damage.

[0031] FIG. 8A is a band energy diagram of hole tunneling layer 114, such as a silicon-nitride-silicon-oxide region designated as N—O in the figure. When the electric field is low, such as during a charge retention state (V_g≈0) or a reading state (V_g≈+1-+2 volts) in some examples, band offset of hole tunneling layer 114 does not occur. Accordingly, tunneling through hole tunneling layer 114 is suppressed, which prevents charge loss from charge storage layer 112, such as a silicon nitride region designed as the second “N” from the left in FIG. 8A. FIG. 8B is a band energy diagram of hole tunneling layer 114, such as a silicon-nitride-silicon-oxide region designated as N—O in the figure. When voltage V_g applied to control gate 116 is non-zero, such as 14 volts, there is a relatively high electric field. Under the high electric field, band offset of hole tunneling layer 114 occurs, thereby providing hole tunneling from control gate 116 to charge storage layer 112, such as silicon a nitride trapping region designed as the second “N” from the left in FIG. 8B.

[0032] The present invention also includes methods of forming NVM cells on a semiconductor substrate. Taking the cell illustrated in FIGS. 1-3 as an example, the method includes forming source region 104 in a portion of substrate 102; forming drain region 106 within a portion of substrate 102, drain region 106 being spaced apart from source region 104; forming well region 108 within a portion of substrate 102, the well region surrounding at least a portion of source and drain regions 104 and 106. In some examples, source, drain, and well regions each is a doped regions and may be formed by doping, implantation, or diffusion techniques or a combination of them.

[0033] The method further includes forming electron tunneling layer 112 over substrate 102; forming charge storage layer 112 over electron tunneling layer 110; forming hole tunneling layer 114 over charge storage layer 112; and forming conductive control gate 116 over hole tunneling layer 114 These layers may be formed by deposition techniques. Additionally, for materials such as silicon nitride or silicon oxide, deposition or a combination of silicon deposition and nitridation or oxidation may be used. And the shape, dimension, and location of each of these layers may be defined by one or more lithography process, which may use a mask to define active regions and apply an etch process to remove regions not needed.

[0034] In this example, electron tunneling layer 110 may provide electron tunneling during a programming operation and may be directly over at least a portion of a region between source and drain regions 104 and 106. Hole tunneling layer 114 may include at least one dielectric layer and may provide hole tunneling during an erasing operation. As noted above, hole tunneling layer 114 may also serve as an electron block layer during at least one of a reading operation and a programming operation. For the layers formed here, the corresponding configuration, location, and materials illustrated above for FIGS. 1-3 may be used, and similar thicknesses may be used in some examples. Additionally, a device formed by the illustrated method may be operated in the ways noted above for the programming, reading, and erasing operations.

[0035] In some examples, an NVM cell may be applied to various memory types including NOR and NAND memories. NVM cells 100 may also be used in a virtual ground array architecture. Additionally, hole tunneling layer 114 may include additional layers of silicon oxide, silicon nitride, or other materials.

[0036] From the foregoing, examples of the present invention are directed to a non-volatile memory semiconductor device having a hole tunneling layer and a method for manufacturing a non-volatile memory semiconductor device having a hole tunneling layer. It will be appreciated by those skilled in the art that changes could be made to the embodiments described above without departing from the broad inventive concept thereof. It is understood, therefore, that this invention is not limited to the particular examples disclosed, but it is intended to cover modifications within the spirit and scope of the present invention as defined by the appended claims.