Static information storage and retrieval – Systems using particular element – Hall effect
Reexamination Certificate
2004-02-10
2004-10-12
Auduong, Gene N. (Department: 2818)
Static information storage and retrieval
Systems using particular element
Hall effect
C365S171000, C365S158000
Reexamination Certificate
active
06804146
ABSTRACT:
FIELD OF THE INVENTION
The invention relates generally to ferromagnetic electronic devices. In particular, the present invention relates to a new nonvolatile hybrid memory element using an improved magnetic spin device in combination with a semiconductor Field Effect Transistor.
BACKGROUND OF THE INVENTION
Solid state Random Access Memories (RAM), presently based on silicon technology, are the most important high-speed, reprogrammable memories in computer systems today. The content of a memory bit is stored in a circuit element called a cell, fabricated in a silicon Integrated Circuit (IC) chip. Each cell can exist in one of two stable states to represent a “0” or a “1.” A word is stored in a particular group of cells in a memory array and can be identified with, and addressed by, a unique numerical address. Thousands to millions of cells can be fabricated in a single Very Large Scale Integrated (VLSI) chip along with additional address decoding and read/write control circuits. Given the address of any word, the memory content of that word can be retrieved during a memory operation time cycle, typically between 5 and 200 nsec. Any given bit can be addressed, written, rewritten, read and reread repeatedly, with fast access, read and write times. The desired attributes of RAM include high speed, low power dissipation, high packing density, and inexpensive manufacturing cost. Although the technology for address decoding, read/write control and read sensing is fairly standard, the kind of circuit element that is used to comprise each memory cell can vary widely and is the subject of much research and development.
Conventional DRAM
The Dynamic Random Access Memory (DRAM) cell is popular because the cell size can be made small, leading to a high packing density and relatively low cost. The storage element is a capacitor C and the two stable states can be, for example, the states with stored charge Q and with stored charge 0. Every cell is connected to an array of write and read wires, also called “bit” and “word” lines. Since one capacitor linked together with other capacitors in an array can lose its charge to its neighbor, the capacitor of each cell is connected to a transistor in that cell so as to isolate it from the array. When the transistor is “on” there is a low resistance to a write or read wire so that an applied voltage can charge the capacitor or a sense circuit can determine the stored charge. When the transistor is “off” there is a high impedance to the write or read wire which isolates the capacitor electrically from any other element in the array.
The packing density of DRAM memories has undergone steady improvement for more than two decades. Early DRAM cells used several capacitors and transistors, but now the most common designs use a single capacitor and single transistor because minimizing the number of elements in a cell allows the size of the cell to be reduced to a minimum. Typically, a Metal Oxide Semiconductor Field Effect Transistor (MOSFET) is fabricated by standard lithographic processing techniques on a silicon substrate. The oxide that isolates the gate from the channel is highly insulating, so that the metallized gate has a capacitance to the rest of the device. In early designs, e.g. with 3 elements per cell, the gate capacitance was used as the storage capacitance. Single element cells use a MOSFET and a separate capacitor C. Reading is performed with a sense circuit that compares the charge (or voltage) of C with the charge (or voltage) of a standard capacitor C in a dummy cell. For an accurate readout, the charge Q stored on the capacitor must be the order of 1 million electrons or more. While FETs have benefited from advances in processing techniques so that they can be fabricated with dimensions smaller than a micron, capacitors have not benefited from similar gains in technology and it is still necessary for capacitors to have dimensions of order one micron or larger in order to hold the necessary charge. Thus, the size of DRAM memory cells is not decreasing as rapidly as processing technology would allow, and one significant drawback of DRAM is that reliable cells might never be fabricated on a submicron scale. Furthermore, the necessity of comparing capacitance with that of dummy cells requires the fabrication of dummy cells which take up extra space on the chip. A second disadvantage, for some cell designs, is that the read process drains whatever charge is stored on the capacitor. This phenomenon is known as “destructive readout” because the read process destroys the state of the cell (capacitor), and a rewrite circuit must be provided to rewrite the memory after every read. The rewrite circuit takes up space on the chip and the rewrite process can lengthen the amount of time of the read cycle. A third weakness is that the capacitor plate of any cell is never perfectly isolated. There is always some finite impedance to ground which permits the charge to leak from the capacitor, and in this process the memory is also lost. To remedy this, a refresh circuit is used to rewrite the memory constantly, typically once every few milliseconds. The refresh circuit takes up space on the chip, uses Central Processing Unit (CPU) time, lengthens the time of the read cycle, and dissipates extra power [typically 0.1 to 0.2 watts per megabyte of RAM]. The power dissipation is of particular concern for memories that are powered by batteries, such as laptop computers, mobile communications equipment, and satellite electronics. DRAM cells derive the name “dynamic” because the memory is dynamic, i.e. it is constantly being refreshed, and memory cells which lose their memory when not being powered are called volatile. A fourth problem with DRAM is that &agr; particle radiation, either from background sources or from contaminants on the chip, can cause a spurious discharge of the capacitor and give a false reading. This is one of the largest error mechanisms limiting the reliability of DRAM.
More recently, cell circuit elements other than typical oxide dielectric capacitors have been proposed to serve as nonvolatile memory storage elements. The chief advantage of a Nonvolatile Random Access Memory (NRAM) is that memory is retained even when the array is not being powered. Thus, power dissipation is minimized since the array draws zero quiescent power. Furthermore, refresh circuitry is eliminated, thus saving space on the chip and requiring less time of the CPU. One category of nonvolatile circuit elements utilizes magnetic materials and their properties. For example, one kind of nonvolatile memory cell uses a circuit element similar to that of DRAM, but the dielectric material of the capacitor is ferroelectric, i.e. the capacitance has two different values for two different states of the dielectric which are determined by application of a magnetic field. Although this cell is nonvolatile, the dielectric properties of the ferroelectric material are relatively weak so that the size of the capacitor C must be fairly large and therefore the cell size is necessarily large. This structure has similar limitations to the oxide dielectric capacitors described above: the read process is destructive, the cells are susceptible to &agr; particle radiation. Finally, the ferroelectric material degrades with time so that the cell is incapable of sustaining an infinite number of read and write cycles.
Other kinds of circuit elements for nonvolatile memory cells use ferromagnetic materials rather than ferroelectric materials. Two different approaches use magnetoresistive elements as the storage elements of the cell.
MRAM
Magnetoresistive Random Access Memory (MRAM) (described in an article by James Daughton, “Magnetoresistive Memory Technology,” Thin Solid Films 216, 162 (1992) which is incorporated by reference herein) employs an array of bit and word lines. Each bit line is divided into n storage cells. Each cell is a trilayer composed of a ferromagnetic metal base layer, a nonmagnetic metal middle layer, and a ferromagnetic metal top layer. The cell has length l, widt
Auduong Gene N.
Gross J. Nicholas
LandOfFree
Hybrid semiconductor—magnetic spin based memory does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Hybrid semiconductor—magnetic spin based memory, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Hybrid semiconductor—magnetic spin based memory will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3284776