Hybrid resistive cross point memory cell arrays and methods...

Static information storage and retrieval – Systems using particular element – Magnetoresistive

Reexamination Certificate

Rate now

  [ 0.00 ] – not rated yet Voters 0   Comments 0

Details

C365S148000

Reexamination Certificate

active

06456524

ABSTRACT:

TECHNICAL FIELD
This invention relates to resistive cross point memory cell arrays and methods of making the same.
BACKGROUND
Many different resistive cross point memory cell arrays have been proposed, including resistive cross point memory cell arrays having magnetic random access memory (MRAM) elements, phase change memory elements, resistive polymer memory elements, polysilicon memory elements, and write-once (e.g., fuse based or anti-fuse based) resistive memory elements.
A typical MRAM storage device, for example, includes an array of memory cells. Word lines may extend along rows of the memory cells, and bit lines may extend along columns of the memory cells. Each memory cell is located at a cross point of a word line and a bit line. Each MRAM memory cell stores a bit of information as an orientation of a magnetization. In particular, the magnetization of each memory cell assumes one of two stable orientations at any given time. These two stable orientations, parallel and anti-parallel, represent logic values of
0
and
1
. The magnetization orientation affects the resistance of a memory cell. For example, the resistance of a memory cell may be a first value, R, if the magnetization orientation is parallel, and the resistance of the memory cell may be increased to a second value, R+&Dgr;R, if the magnetization orientation is changed from parallel to anti-parallel.
In general, the logic state of a resistive cross point memory cell may be read by sensing the resistance state of the selected memory cell. Sensing the resistance state of a single memory cell in the array, however, typically is difficult because all of the memory cells in a resistive cross point memory cell array are interconnected by many parallel paths. Thus, the resistance that is seen at one cross point equals the resistance of the memory cell at that cross point in parallel with resistances of memory cells in the other word lines and bit lines. In addition, if the target memory cell being sensed has a different resistance state due to stored magnetization, a small differential voltage may develop. This small differential voltage may give rise to parasitic or “sneak path” currents that may interfere with the sensing of the resistance state of the target memory cell.
Thus, one hurdle that must be overcome before high density and fast access resistive cross point memories may be developed is the reliable isolation of selected resistive cross point memory cells while data stored on a selected memory cell is being sensed. In general, prior techniques for isolating such memory cells fall into one of three memory cell isolation categories: select transistor isolation techniques; diode isolation techniques; and equipotential isolation techniques.
Select transistor isolation techniques typically involve inserting a select transistor in series with each resistive cross point memory cell. This architecture typically is characterized by fast read access times. Unfortunately, such a series transistor architecture typically also is characterized by relatively poor silicon area utilization because the area under the resistive cross point memory cell array typically is reserved for the series transistors and, therefore, is unavailable for support circuits. In addition, this isolation technique also tends to suffer from relatively poor memory cell layout density because area must be allocated in each memory cell for a via that connects the memory cell to the series transistor in the substrate. This isolation technique also generally requires relatively high write currents because an isolated write conductor must be added to the memory cell to provide a write circuit in parallel with a read circuit and the location of the write conductor results in high write currents to generate the required write fields. In general, this approach is limited to a single memory plane because the series transistors must be located in the substrate and there is no practical way to move the series transistors out of the substrate and into the memory cell plane.
Diode isolation techniques typically involve inserting a diode in series with each resistive cross point memory element. This memory cell array architecture may be implemented with thin film diodes that allow multi-level resistive cross point memory arrays to be constructed (see, e.g., U.S. Pat. No. 5,793,697). This architecture has potential for high-speed operation. The difficulty often associated with this architecture involves providing a suitable thin film diode with minimum process feature sizes matching the potential density of the memory cell arrays. In addition, this approach uses one diode per memory element and, at currently practical MRAM features and parameters, for example, each diode would be required to conduct 5 to 15 kA/cm
2
. Such high current densities generally are impractical for implementing thin film diodes in high-density MRAM arrays.
Equipotential isolation techniques typically involve sensing resistive cross point memory cells without using series diodes or transistors (see, e.g., U.S. Pat. No. 6,259,644). This approach may be implemented by a cross point array of memory elements that is relatively simple to fabricate. This cross point memory cell array architecture typically has a density that is limited only by the minimum feature sizes of the implementing circuit technology and typically requires relatively low write currents. In addition, it is relatively simple to extend this approach to multi: level resistive cross point memory cell arrays to achieve very high-density memories. Equipotential isolation, however, often is difficult to implement in large arrays. Auto-calibration and triple sample read techniques have been used to sense data in large MRAM arrays using equipotential isolation techniques, but these sense processes typically limit the read sense time to a range of 5 &mgr;s to 20 &mgr;s.
SUMMARY
The invention features a data storage device that includes a novel resistive cross point memory cell array that enables high-density fabrication and high-speed operation with isolation diodes that have practical dimensions and current density characteristics. In addition, the inventive data storage device includes a novel equipotential isolation circuit that substantially avoids parasitic currents that otherwise might interfere with the sensing of the resistance state of the memory cells.
In one aspect, the invention features a data storage device that includes a resistive cross point array of memory cells, a plurality of word lines, and a plurality of bit lines. The memory cells are arranged into multiple groups of two or more memory cells. The memory cells of each group are connected between a respective word line and a common isolation diode that is coupled to a bit line.
Embodiments of the invention may include one or more of the following features.
Multiple read circuits preferably are each coupled to one or more associated groups of memory cells by a respective bit line. The read circuits preferably are operable to sense current flow through a memory cell of the associated groups. Each read circuit may include a differential amplifier. The differential amplifier may be a current mode differential amplifier. The differential amplifier preferably is operable to compare current flowing through a selected memory cell with current flowing through one or more reference cells. The data storage device may further comprise multiple comparator circuits each of which is coupled to an associated read circuit. The comparator circuits preferably are operable to convert an analog differential sense voltage to a digital output read signal.
The data storage device preferably comprises an equipotential generator that is coupled to the word lines and the bit lines. The equipotential generator preferably is operable to set voltage levels in the resistive cross point memory cell array to substantially prevent parasitic currents from flowing through unselected memory cells. The equipotential generator may be operable to set an input node of the common isolation

LandOfFree

Say what you really think

Search LandOfFree.com for the USA inventors and patents. Rate them and share your experience with other people.

Rating

Hybrid resistive cross point memory cell arrays and methods... 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 resistive cross point memory cell arrays and methods..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Hybrid resistive cross point memory cell arrays and methods... will most certainly appreciate the feedback.

Rate now

     

Profile ID: LFUS-PAI-O-2832639

  Search
All data on this website is collected from public sources. Our data reflects the most accurate information available at the time of publication.