Patent Application
20070018245
The present invention relates to a semiconductor device, and more specifically, to a nonvolatile memory with FinFET structure capable of trapping carriers.
The semiconductor industry has been advanced to the field of Nanometer Scale technologies. The fabrication of the nonvolatile memories also follows the trend of the size reduction. The nonvolatile memory includes various types of devices which have been developed for specific applications. One of the applications of flash memory is BIOS for computers. The high-density nonvolatile memories can be applied as the mass storage of portable handy terminals, solid-state camera and PC cards. It is because that the nonvolatile memories exhibit many advantages, such as memory retention without power, fast access time, low power dissipation in operation, and robustness.
The formation of nonvolatile memories toward the trends of low supply power and fast access, because these requirements are necessary for the application of the mobile computing system. Nonvolatile memory devices include a floating gate to store charges and an element for electrically placing charge in and removing the charges from the floating gate. At present, the low voltage nonvolatile memory is applied with a voltage of about 3V to 5V during charging or discharging the floating gate. In order to attain high tunneling efficiency, the thickness of the dielectric between the floating gate and substrate has to be scaled down due to the supply voltage is reduced. The data program method of a non-volatile memory device includes a method using Fowler-Nordheim (FN) tunneling. In FN tunneling, a high voltage is applied to a control gate to induce a high electric field in a tunnel oxide layer, and electrons of a semiconductor substrate pass the tunnel oxide layer and are injected into a floating gate. During the mode of erasing, the bias may apply on the source to discharge the electron from the floating gate to the source of a memory device.
Currently, the SOC (system on chip) desires memory with high operation speed and integrated in one single chip. For example, the single polysilicon processing may integrate with other devices such as transistors. The typical non-volatile memory employs stack gate memories by double polysilicon processing. One type of the memories uses trapping layer instead of floating gate (FG) to hold the carrier. The memory cells are constructed with a trapping ONO or ON layer. The memory cell generally includes a P-type silicon substrate and two PN junctions between N+ source (or drain regions) and P type substrate, a nitride layer sandwiched between two oxide layers and a polycrystalline layer. To program or write the cell, voltages are applied to the drain and the gate and the source is grounded. These voltages generate an electric field along the length of the channel from the source to the drain. This electric field causes electrons to be drawn off the source and begin accelerating towards the drain. The hot electrons are generated at the boundary between drain and channel during the acceleration.
In the prior art, please refer to U.S. Pat. No. 4,881,108 U.S. Pat. No. 5,768,192 to Eitan B. entitled “Non-volatile semiconductor memory cell utilizing asymmetrical charge trapping”, filed on 16 Jun. 1998. The charge trapping memory may also be referred to U.S. Pat. No. 6,335,554 to Yoshikawau and Kuniyoshi, entitled “Semiconductor Memory”. The patent disclosed a memory with ONO structure. Further article teaches the memory with ONO stacked layer could also be found. Please refer to the article, Chan, T. Y. et al, “A True Single-Transistor Oxide-Nitride-Oxide EEPROM Device,” IEEE Electron Device Letters, vol. EDL-8. No. 3, March 1987.
Furthermore, another type structure called FinFET is developed recently. The channel region of the device is similar to a Fin structure that is why it was named as FinFET device. The source/drain and the channel are formed in a single crystallographic orientation silicon layer. Gate oxide is subsequently formed over the silicon layer, followed by forming a gate wrapping the channel region, thereby creating the creating the dual gate structure. Thus, the channel width of the FinFET device is wider than the prior structure. It allows generating higher driving current than that of the conventional FET device. In the prior art, please refer to U.S. Pat. No. 4,881,108 U.S. Pat. No. 6,800,910 to Lin; Ming-Ren, and assigned to Advanced Micro Devices, Inc., entitled “FinFET device incorporating strained silicon in the channel region”. The filing date is Dec. 31, 2002. Other reference could refer to U.S. Pat. No. 6,770,516, assigned to Taiwan Semiconductor Manufacturing Company, entitled “Method of forming an N channel and P channel FINFET device on the same semiconductor substrate”. The filing date is Sep. 5, 2002.
The object of the present invention is to disclose a Fringing Field Induced Localized Charge Trapping Memory.
The present invention discloses n-MOSFET with gate-to-drain non-overlapped implantation (NOI) as a non-volatile memory device capable of storing multi-bits per cell.
The present invention discloses a fringing field induced charge trapping memory comprising: an isolation or semiconductor substrate having a semiconductor layer formed thereon. The semiconductor layer includes a planar or non-planar channel. Source and drain regions are formed in a portion of the semiconductor layer, wherein a pocket ion implantation region is optionally formed adjacent to the source and drain regions, the conductive type of the pocket ion implantation region is opposite to the one of the doped region. A gate dielectric layer is formed on the semiconductor layer and a gate structure is next formed on the gate dielectric layer. The gate structure of includes dual gate, triple gate or FinFET structure. A normal field induced channel is formed in the semiconductor layer and under the gate structure. A multi-portion structure consisting of a perpendicular portion and a parallel portion is formed on the sidewall of the gate structure and on the semiconductor layer, wherein the parallel portion acts as tunneling part and the perpendicular portion acts as a charge barrier. The perpendicular portion and the parallel portion are formed simultaneously or respectively.
A fringing field induced channel is formed in the semiconductor layer and under the parallel portion of the multi-portion structure. A charging (or spacer) for charge trapping is formed on the multi-portion structure for storing carriers. Metal-semiconductor compound or silicide is optionally formed on the gate structure and the source and drain regions. One aspect of the present invention is that a Non-Overlapped Implantation structure is formed by the gate structure and the source and drain regions. The configuration of the gate structure and the source, drain regions is substantially orthogonal.
The fringing field induced channel is formed between the normal field induced channel and the drain region. The carrier flows from the normal field induced channel to the drain region through the fringing field induced channel during a programming mode. The hot carriers are injected from the fringing field induced channel and stored in the charging layer through the multi-portion structure. In a read mode, carrier flows from the drain region to the normal field induced channel through the fringing field induced channel.
In one embodiment, the memory further comprises lightly doped drain region formed adjacent to the source and drain regions, wherein the junction of the lightly doped drain region being shallower than the one of the source and drain regions and under the gate structure. A double doped drain region could also be formed adjacent to the source and drain regions, wherein the junction of the double doped drain region being deeper than the one of the source and drain regions and under the gate structure.
The charging layer (or spacer) is formed of nano-scale particles or nitride or the material having energy gap lower than 6 eV The memory further comprises a first dielectric layer (or spacer) formed on the charging layer (or spacer) wherein the first dielectric layer (or spacer) includes oxide or the material having energy gap greater than 7 eV. Similarly, the memory could also comprises a second dielectric layer (or spacer) formed on the first dielectric layer (or spacer) wherein the second dielectric layer (or spacer) includes oxide or nitride.
The material of the gate dielectric layer includes silicon oxide (SiO2), zirconium oxide (ZrO2) or hafnium oxide (HfO2). The silicide includes TiSi2, CoSi2 or NiSi.
The foregoing aspects and many of the attendant advantages of this invention will become more readily appreciated as the same becomes better understood by reference to the following detailed description, when taken in conjunction with the accompanying drawings, wherein:
The present invention proposes a novel structure of a nonvolatile memory for charge trapping. In the structure, the cell capacity for storing data can be increased by the cell structure. One aspect of the present invention is the gate-to-source/drain non-overlapped implantation (NOI). The FinFET device according to the invention allows storing single or multiple bits in a cell.
Please refer to
The detail description will be seen as follows. A substrate 2 is provided for the present invention. In a preferred embodiment, as shown in the
The source and drain region 14 is formed by performing an ion implantation to dope ions into the layer 2 using the gate structure 6 and sidewall spacers 12 as a implanting mask. Silicide 16 is introduced on the exposed surface of the top portion of gate and the semiconductor layer 2 on the source and drain regions 14 to reduce their resistance. Preferably, the silicide 16 can be TiSi2, WSi2, CoSi2 or NiSi. The gate structure 6 acts as the control gate, and the nitride spacers are used to trap carriers. The number 8 in the figure indicates the charge that is trapped. The spacers 12 are used to store charges, thereby defining the digital states including (0, 0), (0, 1), (1, 0), (1, 1). A sectional view of a multi-bit nonvolatile memory cell in accordance with the present invention is shown in drawings. The memory cell includes a semiconductor layer 2 having at least two junctions, one is the left junction and the other is the right junction. Channels are located between the two junctions during operation. Above the normal field channel is an oxide 4, on top of the oxide layer 4 is a control gate 6. Spacer 12 is used for charge trapping and is preferably comprised of silicon nitride. The hot electrons or holes are trapped as they are injected into the Spacer 12. The memory cell is capable of storing two bits of data, a right bit and a left bit. The two-bit memory cell is a symmetrical device. The left junction serves as the source terminal and the right junction serves as the drain terminal for the right bit programming. Similarly, for the left bit programming, the right junction serves as the source terminal and the left junction serves as the drain terminal. When the distinction between left and right bits is not crucial to the particular discussion, the terms source and drain are utilized as conventional manner.
The electric field adjacent to the S/D is significant due to the P-N junctions formed by the NOI structure. The hot carrier induced by the NOI occurs at the region right under the sidewall spacers 12 and the hot carrier will inject into the spacers 12 through the tunneling dielectric under the fringing induced field. On the other hand, while the dosage of the source/drain extension portion or lightly doped drain (LDD) is reduced or omitted, the structure could increase the hot carrier current during the write mode and gains the higher writing speed. In order to adjust the writing speed, the LDD structure could be removed or omitted. Based on the drain junction is generated under the nitride spacers 12, the channel hot carriers are expected to be generated at the edge of the drain junction, thereby injecting the carriers into the nitride spacers 12.
Turning to
Alternatively, the other embodiments shown in
Please refer to
Similarly, the preferred embodiments from
Next, the embodiments shown in
In the illustrations of the present invention, form the planar parallel to the surface of the paper, two spacers are located adjacent to the gate, hence, two-bits may be stored in the nitride spacers. Therefore, the present invention provides a multi-bit device for storing multi-bits digital signal. The trapping layer is formed of ONO or ON configuration to trap the carrier, thereby defining the digital states.
The electric field near the S/D junction becomes significant due to the abrupt junction formation of NOI process. The NOI-induced hot carrier effect takes place right under the SiN spacer. The drain node can be defined as first bit and source is defined as second bit, vice verse. The programming is achieved by tunneling effect. The read operation is obtained by applying reverse bias to the source and grounding the drain. During the operation of the NOI n-MOSFETs, the channel current is generated at the drain and it is not significant at the high electronic field. In lieu of altering the current direction between the S/D by the mechanism of reverse read mechanism, it is possible to sense the status of the programming region. In one embodiment, during the reading mode, the S/D bias is 1V/0V, the threshold difference between two bits is higher than 1.2 V. Based on the NOI structure formed between the gate and S/D junction, the induced fringing field is extremely low compared to the conventional SONOS. Therefore, the NOI n-MOSFETs's low induced fringing field and low drain current will lead to longer life time for product.
As will be understood by persons skilled in the art, the foregoing preferred embodiment of the present invention is illustrative of the present invention rather than limiting the present invention. Having described the invention in connection with a preferred embodiment, modification will now suggest itself to those skilled in the art. Thus, the invention is not to be limited to this embodiment, but rather the invention is intended to cover various modifications and similar arrangements included within the spirit and scope of the appended claims, the scope of which should be accorded the broadest interpretation so as to encompass all such modifications and similar structures. While the preferred embodiment of the invention has been illustrated and described, it will be appreciated that various changes can be made therein without departing from the spirit and scope of the invention.
