Development of semiconductor memory devices has been increasing at a fast pace because of major breakthroughs in materials, manufacturing processes and designs of semiconductor devices. Semiconductor device manufacturers are constantly enhancing their efforts for more advanced miniaturization, high-integration and capacity increase of the semiconductor devices. This has given an impetus to research and development for more stabilization, higher speeds and smoother operation of semiconductor devices. These results have been brought about by the device makers improving the process techniques, microminiature device techniques and circuit design techniques in the manufacture of semiconductor memory cells such as SRAM (Static Random Access Memory).
The push for ever increasing device densities is particularly strong in CMOS (complementary metal oxide semiconductor) technologies, such as the in the design and fabrication of field effect transistors (FETs). Unfortunately, increased device density in CMOS FET often results in degradation of performance and/or reliability.
Multiple port SRAM cells have been implemented in planar CMOS bulk processes. However, such bulk processes do not exhibit a desirable sub-threshold slope, and exhibit matching and noise immunity difficulties.
Dual port CMOS SRAM are capable of performing read and write operations at high speeds. In general, one unit memory cell of a single port CMOS SRAM device is composed of six transistors, that is, two access transistors, and four transistors configured as an inverting latch to perform the read and write operations sequentially. Word lines are coupled to the access transistors and data is provided or read on bit lines. In contrast, a dual port CMOS SRAM device is configured with an addition of two access transistors coupled to an additional word line, and a pair of additional bit lines so as to perform two read operations in parallel.
In a read operation, an externally received read address signal is decoded, and according to the decoding result, a word line signal for the read operation is enabled. Next, the access transistors are turned on, and the data stored at the latch is read through the bit line and the complementary bit line. Similarly, in the write operation, a write address signal is received and is decoded, and according to the decoding result, a word line signal for a write operation is enabled. The access transistors are then turned on, and the data loaded on the bit line and the complementary bit line is stored at the latch.
In the following description, reference is made to the accompanying drawings that form a part hereof, and in which is shown by way of illustration specific embodiments which may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention, and it is to be understood that other embodiments may be utilized and that structural, logical and electrical changes may be made without departing from the scope of the present invention. The following description of example embodiments is, therefore, not to be taken in a limited sense, and the scope of the present invention is defined by the appended claims.
A multiple port SRAM (static random access memory) cell is formed with MuGFETs (multi gate field effect transistors). The MuGFET based multiple port SRAM cell may have a better sub-threshold slope, improved matching and better noise immunity than provided by typical CMOS (complementary metal oxide semiconductor) bulk processes. The use of MuGFET transistors may also yield a compact memory cell with excellent leakage characteristics, as the off current flowing through access devices is significantly lower compared to common bulk CMOS devices. An example dual port circuit and layout are shown, along with an example triple port and n-port circuits.
In one embodiment, the transistors are a MuGFET type of transistor, with the pull down transistors 130, 132 having one fin, or more fins for increased current capacity. The gate of the MuGFET transistor traverses at least both sides of the fin to provide a multiple gate effect on the fin, which serves as the channel of the device.
The pull up transistors may be p-type in one embodiment, with the other transistors n-type. In a further embodiment, the pull up transistors may be n-type, again with the other transistors p-type.
In one embodiment, a second set of access transistors 140, 142 are coupled between the pull up and pull down transistors in the same manner as the first set of access transistors 110 and 112. Their gates are coupled to a word line B at 150 with sources coupled to bit line B at 155 and 157.
In operation, the complementary bit lines may be charged and the various sets of access transistors may be turned on by their respective word lines. Sense amplifiers may then be used to sense the values on the bit lines to determine the value stored in the memory cell. One or both of the access transistor pairs may be turned on simultaneously or asynchronously along with the bit line charging to provide separate access to the memory cell by different devices or by different portions of a same device.
The fin 210 has a top surface 250 and laterally opposite sidewalls 255. The semiconductor fin has a height or thickness equal to T and a width equal to W. The gate width of a single fin MuGFET transistor is equal to the sum of the gate widths of each of the three gates formed on the semiconductor body, or, T+W+T, which provides high gain. If MuGFET devices are formed on an insulator, better noise immunity may result. Formation on the insulator provides isolation between devices, and hence the better noise immunity. Better noise immunity allows the formation of multiple ports without interference with other transistors in an SRAM. Having the gate traverse two or more sides of the fin or channel results in much quicker off current than prior bulk CMOS devices.
The use of MuGFET transistors may also provide a better subthreshold slope that is steeper than in bulk CMOS devices, so the device switches off more quickly. Since the channels are formed by the use of narrow fins, improved matching of the devices may be significantly easier than in planar bulk CMOS devices, allowing better control of their current characteristics.
Many other layouts may be used by those of skill in the art to form the dual-port MuGFET SRAM core cell of
Additional sets of access devices, word lines and bit lines may be provided in further embodiments to provide even further ports.
The Abstract is provided to comply with 37 C.F.R. §1.72(b) to allow the reader to quickly ascertain the nature and gist of the technical disclosure. The Abstract is submitted with the understanding that it will not be used to interpret or limit the scope or meaning of the claims.