Patent Application
20060046354
The present invention relates to a semiconductor device, and more particularly, to a recessed gate dielectric antifuse.
A conventional Metal Oxide Semiconductor Field Effect Transistor (MOSFET) comprises a doped polysilicon gate, a channel conduction region, and source/drain regions formed by diffusion of dopants in the silicon substrate. The channel region separates the source from the drain in the lateral direction. A substantially planar layer of a dielectric material that prevents electrical current flow separates the polysilicon gate from the channel. The substantially planar dielectric material separating the polysilicon gate from the channel region, henceforth referred to as the gate oxide, usually consists of the thermally grown silicon dioxide (SiO2) material that leaks very little current through a mechanism, which is called Fowler-Nordheim tunneling under voltage stress.
A reliable, low resistance, programmable antifuse may be formed from such a conventional MOSFET, which provides a high impedance between the gate and the source, the drain, the substrate, or a well. When the MOSFET is stressed beyond a critical electrical field by applying sufficient voltage to the gate, the transistor is destroyed by rupturing the gate oxide. In particular, the gate oxide over the substrate becomes a resistive short, causing part of the gate oxide to form current paths by diffusion of polysilicon gate material or silicon from the substrate.
To increase device yield, semiconductor integrated circuits such as EPROMs, flash EEPROMs, DRAM, SRAM, and other various random access memory types employ redundant circuitry that allows the integrated circuits to function despite the presence of one or more manufacturing or other defects by employing the redundant circuitry rather than the original, defective circuitry. Such conventional memories often use gate-oxide antifuses as part of their redundancy scheme wherein the redundant circuitry may be employed in place of defective circuitry by blowing one or more of the antifuses.
To keep pace with the current trend toward maximizing the number of circuit devices contained in a single chip, integrated circuit designers continue to design IC devices with smaller and smaller feature sizes.
It is against the above mentioned background, that the present invention provides a number of unobvious advantages and advances over the prior art. In particular, the present invention discloses a recessed dielectric antifuse device and method of formation thereof. The antifuse device includes a substrate and laterally spaced source and drain regions formed in the substrate. A recess is formed between the source and drain regions. A gate and gate oxide are formed in the recess and lightly doped source and drain extension regions contiguous with the laterally spaced source and drain regions are optionally formed adjacent the recess. Programming of the recessed dielectric antifuse is performed by application of power to the gate and at least one of the source region and the drain region to breakdown the dielectric, which minimizes resistance between the gate and the channel.
The aforementioned features and advantages of the present invention, as well as additional features and advantages thereof, will be more clearly understood hereinafter as a result of a detailed description of the various embodiments of the invention when taken in conjunction with the following drawings, wherein like elements are indicated by like symbols.
Skilled artisans appreciate that elements in the drawing are illustrated for simplicity and clarity and have not necessarily been drawn to scale. For example, the dimensions of some of the elements in the drawing may be exaggerated relative to other elements to help to improve understanding of the various embodiments of the present invention.
The invention disclosed herein is directed to a recessed gate oxide antifuse and a method of fabricating the recessed gate oxide antifuse structure. The drawing figures illustrate a partially completed integrated circuit device. In the following description, numerous details are set forth in order to provide a thorough understanding of the present invention. It will be appreciated by one skilled in the art that variations of these specific details are possible while still achieving the results of the present invention.
Additionally, well-known processing steps are not described in detail in order not to unnecessarily obscure the present invention. For the sake of brevity, conventional electronics, semiconductor manufacturing, memory technologies and other functional aspects of the devices (and components of the individual operating components of the devices) may not be described in detail herein. Furthermore, for purposes of brevity, the invention is frequently described herein as pertaining to an antifuse for use in electrical or electronic systems.
It should be appreciated that many other manufacturing techniques could be used to create the antifuses described herein, and that the techniques described herein could be used as individual devices, grouped together in discrete circuits, formed in memory arrays, or integrated with other devices. Further, the techniques would be suitable for application in electrical systems, optical systems, consumer electronics, industrial electronics, wireless systems, appliances, space applications, or any other application.
The term antifuse and the terms storage or programmable coupled with the terms cell, element, or device are often used interchangeably in this field. The present invention is applicable to all these terms as they are generally understood in the relevant art.
The terms chip, integrated circuit, monolithic device, semiconductor device and microelectronic device are often used interchangeably in this field. The present invention is applicable to all of these terns as they are generally understood in the field.
The terms pins, pads and leads refer to input and/or output terminals of a connector, device, chip, printed circuit, or the like, which are used to provide electrical connection to one or more connectors, devices, chips, printed circuits, or the like. The present invention is applicable to all of these terms as they are generally understood in the field.
The terms metal line, trace, wire, conductor, signal path and signaling medium are all related. These related terms are generally interchangeable and appear in order from most specific to most general. In this field, metal lines are sometimes referred to as traces, wires, lines, interconnect or simply metal. Metal lines, generally aluminum (Al) or an alloy of Al and copper (Cu), are conductors which provide signal paths for coupling, or interconnecting, electrical circuitry. Conductors other than metal are available in microelectronic devices. Materials such as doped polysilicon, doped single-crystal silicon (often referred to simply as diffusion, regardless of whether such doping is achieved by thermal diffusion or ion implantation), titanium (Ti), molybdenum (Mo), or refractory metal silicides are examples of other conductors. Signaling medium is the most general term and encompasses the others.
Moreover, it should be understood that the spatial descriptions (e.g., “above”, “below”, “up”, “down”, “top”, “bottom”, “beneath”, “across”, etc.) made herein are for purposes of illustration only, and that practical antifuses can be spatially arranged in any orientation or manner. Arrays of these antifuses can also be formed by connecting them in appropriate ways and with appropriate devices.
As discussed below, the following sections more fully describe the present invention.
A cross-sectional view of a conventional MOSFET device 100 is shown in
Within the substrate 114 are formed heavily doped source and drain regions 116 and 118, and lightly-doped source and drain extension regions 120 and 122, respectively. Generally, doped regions are regions containing a higher concentration of p-type or n-type dopants than the rest of the substrate 114. The source and drain extension regions 120 and 122 generally have a lower concentration of dopants compared to the source and drain regions 116 and 118. However, the regions 116, 118, 120 and 122 can be doped to the same levels. The region in the substrate directly below the gate 112, and between source and drain extension regions 120 and 122, is typically referred to as a channel region 124.
Resistive current paths can be formed between the gate 112, the channel region 124, or the source and drain extension regions 120 and 122. The lowest resistance paths tend to be between the gate 112 and the source or drain regions 116 and 118 and the source or drain extension regions 120 and 122. However, when the resistive current paths are formed from the gate electrode 112 to the substrate 114, which can comprise a P-well or N-well, the fuse resistance is much higher. It is known by those having ordinary skilled in the art that soft breakdown (i.e., high fuse resistance) and hard breakdown (i.e., low fuse resistance) are influenced by a breakdown spot position. The soft breakdowns almost exclusively occur between the gate 112 and channel region 124 (i.e., substrate, N-well, or P-well).
Both high resistance and low resistance breakdown paths are illustrated in
For antifuse applications, the soft breakdowns are undesirable because fuse resistance distribution is considerably higher than that of the hard breakdown. Also, the fuse resistance of the conventional MOSFET antifuse device initially decreases in channel length, but begins to increase with a further decrease in channel length for a given power supply applied to the antifuse. Furthermore, the variability of the fuse resistance increases with a decrease in channel length. In particular technologies, in order to achieve a high density memory cell, the fuse structure should be as small as possible. Thus, conventional MOSFET antifuse devices are not suitable to provide for a small redundancy or error correction structure of a high density memory cell due to variability of the fuse resistance and increase in fuse resistance.
The present invention is directed to a device and associated manufacturing method that overcome these above-noted deficiencies in conventional MOSFET antifuses.
After defining shallow trench isolation (STI) regions, the substrate 210 is patterned to form recess regions in the active areas, such as for example, a recess 212 formed in the substrate 210. Width of the recess 212 is about 0.5 times the wordline width. Depth of recess 212 is adjusted to optimize electrical characteristics of the access device. Next a sacrificial oxide/wet etch oxide 213 for surface smoothing is grown in the recess 212. This sacrificial oxidation is optimized to minimize a thickness delta between the bottom and the sidewall of the recess 212.
A gate dielectric layer 214 is formed over the surface of substrate 210, lining the recess 212, using conventional semiconductor processing techniques, thereby forming a recessed gate dielectric. In one embodiment, the gate dielectric layer 214 is an oxide, such as for example, silicon dioxide, and has a thickness of between about 20 and about 50 angstroms. In one embodiment, the gate oxide 214 is grown with the oxidation again being optimized to minimize a thickness delta between the bottom and the sidewall of the recess 212. The gate oxide 214 is then patterned (e.g., photo/etch) as needed to define thin/thick oxide regions.
The present invention, however, is not limited to antifuses having oxide dielectrics. Those of ordinary skill in the art will be able to substitute other known antifuse dielectric materials into the antifuse structure disclosed herein. For example, in other embodiments, amorphous silicon, silicon nitride, ONO, tantalum oxide (Ta2O5), BST, PZT, lanthanide materials, and combinations thereof may be a suitable dielectric. Additionally, the dielectric thickness will, in part, determine the voltage at which programming of the antifuse element will be achieved. Proper selection of the thickness will assure both unprogrammed antifuse integrity and that the program voltage will be low enough such that an integrated circuit process can be used to build an array using the antifuse element.
An optional second gate oxidation may be performed to provide a desired gate oxide 214 thickness for access and periphery thick oxide devices. Once again, the second gate oxidation is optimized to minimize thickness delta between bottom of recess and sidewall of recess.
Next, a gate 216 is formed on the surface of dielectric layer 214, where by the thickness of the deposition completely fills the recess 212. The gate 216 is doped, in-situ or via implantation, to provide poly dopants (n+poly, p+poly), a desired threshold voltage, sheet resistance, and the like. In one embodiment, gate 216 is an amorphous silicon having a thickness of about between 1,500 and 3,000 angstroms, and doped to a sheet resistance of about 30 ohms/square. The gate 216 is defined using conventional processing techniques where by the remaining exposed gate oxide may be removed using standard semiconductor etching techniques. In other embodiments, the gate 216 may be metal, silicon, polysilicon, polycide, silicide, and combinations thereof. Rapid thermal processing (RTP) is then performed to active the poly dopants.
After gate 216 is defined, optionally, first and second lightly doped extension regions 218 and 220 of a second conductivity type are formed in substrate 210 using the well-known self-aligned gate process. In one embodiment, the first and second regions 218 and 220 are lightly doped source and drain (LDD) regions formed in the substrate 210 adjacent the channel region. In the illustrated embodiment, the lightly doped drain (LDD) are n− regions formed by implanting ions, selected from phosphorus, arsenic, and antimony ions, at energies within ranges that are conventional in the art to achieve a selected source and drain impurity ion amount. Typically, ions are implanted with dosages of from between about 1×1013 to about 1×104 atoms/cm2 at energies of between about 5 to 80 KeV at a vertical angle to provide an average dopant concentration for the LDD regions ranging from about 1×1017 ions/cm3 to 1×1019 ions/cm3. For LDD p− regions, boron, boron bifluoride (BF2), or borane (B2H10) ions are used with the same dosages and energies, and resulting in the same concentrations as mentioned above.
Spacers 222 and 224 are then formed on the sidewalls of gate 216 and in one embodiment, are formed by providing a layer of CVD SiO2 and etching it back, leaving spacers 222 and 224. In other embodiments, the spaces 222 and 224 may be silicon nitride, or any other suitable spacer material. The spacers 222 and 224 may have a base width from about 100 to about 500 Angstrom. In another embodiment, an insulating capping layer 225, integral with spacers 222 and 224 may formed and having a thickness of between from about 100 to about 200 Angstroms. The capping layer 225 may be the same material as the spacers 222 and 224, or any other suitable insulating material.
Third and fourth regions 226 and 228 of the second conductivity type are then formed in the silicon substrate 210 using known methods. In one embodiment, regions 226 and 228 are heavily doped source and drain (HDD) regions formed by the ion implantation of n+ ions. Ions selected from phosphorus, arsenic, antimony ions and the like are implanted at energies within ranges that are conventional in the art to achieve a selected source and drain impurity ion amount. Typically, n+ ions are implanted at dosages of between about 1×1014 to about 8×1015 atoms/cm2, and at energies of between about 5 to about 80 KeV at a vertical angle to provide an average dopant concentration for the diffusion regions ranging from about 1×1017 ions/cm3 to 1×1019 ions/cm3. For HDD p+ regions, boron, boron bifluoride (BF2), or borane (B2H10) ions are used with the same dosages and energies, and resulting in the same concentrations as mentioned above. Regions 226 and 228 are self aligned to the edges of spacers 222 and 224. Regions 226 and 228 are so located that region 226 is contiguous with region 218 and region 228 is contiguous with region 220.
Next, an insulating region 230 may be formed. Contact holes are then etched in insulating region 230 and contacts 232 and 234 are formed to regions 226 and 228 and contact 236 is formed to gate 216. Contacts 232, 234 and 236 are used to form electrical connections to a metal layer for connection to other circuitry in the integrated circuit as is well known in the art. In one embodiments, the contacts 232, 234, and 236 may be a metal, titanium tungsten (TiW), polysilicon, and combination thereof.
Before programming, antifuse 200 is an open circuit, the resistance between gate 216 and the regions 218, 220, 226, and 228 in the silicon substrate 210 being higher than 1×109 ohms. A low resistance filament 238 may be formed between regions 220 and/or 228 and the gate 216 by applying a programming voltage in the range of 6-8 volts on region 228 with respect to the substrate, with region 226 held at the substrate potential while the gate 216 is grounded or biased at a slightly positive voltage, i.e., a voltage in the approximate range of 0-2 volts, with respect to the substrate. All voltages are measured with respect to the substrate.
Under these conditions, the device will be brought into snap-back breakdown. Snap-back breakdown is a well known phenomenon, characterized in the structure of
Under the snap-back breakdown condition, dielectric 214 in the antifuse 200 can be broken down in milliseconds or a shorter time. After breakdown, contact 236 will be electrically connected to contact 234 through gate 216, with an ohmic connection through the rupture in dielectric 214 located over regions 220 and/or 228. Thus, the antifuse is programmed.
It is to be appreciated that regions 218 and 220 are optional. Accordingly, if the structure disclosed with respect to
Referring now to
Field oxide regions separate the pair of antifuses 310a and 310b from other antifuse pairs on the same row as is conventionally employed in integrated circuit layout design. Antifuses 310a and 310b share a common source 326 and are each identical to the antifuse described with reference to
As shown in
As illustrated by
An advantage of the present invention is the compact nature of nonvolatile one-time programming antifuse devices, which can be manufactured using standard 0.13 μm or other CMOS processes. Additionally, the present invention with its recessed gate dielectric arrangement, provides in one embodiment about a 20 percent reduction in gate oxide thickness, and about a 25 percent reduction in gate thickness as compared to conventional gate-oxide antifuses. Furthermore, the present invention provide a longer gate length in a smaller horizontal area, such that programming can use a lower voltage than the conventional programming voltages of 10 to 12 volts used to rupture the gate oxide. Moreover, the antifuse devices of the present invention minimize resistance variances between blown antifuse devices and are optimized with small chip area utilization. Thus, integrating multitudes of antifuses on a single IC can be achieved according to the present invention.
The present invention can be implemented with various changes and substitutions to the illustrated embodiments. For example, the present invention can be useful in programmable application-specific integrated circuits (ASICs), such as programmable logic devices (PLDs) and field-programmable gate arrays (FPGAs).
The above described embodiments are intended to illustrate the principles of the invention, not to limit its scope. Other embodiments in variations to these preferred embodiments will be apparent to those skilled in the art and may be made without departing from the spirit and scope of the invention as defined in the following claims.
