1
MAGNETOELECTRONIC MEMORY
ELEMENT WITH INDUCTIVELY COUPLED
WRITE WIRES
RELATED APPLICATION DATA 5
The present invention claims priority to and is a continuation of application Ser. No. 10/853,791 filed May 24, 2004 (now U.S. Pat. No. 6,888,746) which is a continuation of application Ser. No. 10/100,210 filed Mar. 18, 2002 titled 10 "Magnetoelectronic Memory Element With Inductively Coupled Write Wires" (now U.S. Pat. No. 6,741,494). Application serial No. 10/100,210 is a continuation of an application Ser. No. 09/532,706 filed Mar. 22, 2000 titled "Magnetoelectronic Memory Element With Isolation Element" 15 (now U.S. Pat. No. 6,388,916). The latter application Ser. No. 09/532,076 is in turn a divisional application of Ser. No. 08/806,028 filed Feb. 24, 1997 entitled "Hybrid Hall Effect Memory Device & Method of Operation," (now U.S. Pat. No. 6,064,083). Ser. No. 08/806,028 is a continuation-in- 20 part of Ser. No. 08/643,805, filed May 6,1996 titled "Hybrid Hall Effect Device and Method of Operation," (now U.S. Pat. No. 5,652,445), which in turn is a continuation-in-part of an application Ser. No. 08/493,815, filed Jun. 22, 1995 titled "Magnetic Spin Transistor Hybrid Circuit Element," 25 (now U.S. Pat. No. 5,565,695); and said Ser. No. 08/806,028 is also a continuation-in-part of an application Ser. No. 08/425,884, filed Apr. 21, 1995 titled "Magnetic Spin Transistor, Logic Gate & Method of Operation," (now U.S. Pat. No. 5,629,549); and an application Ser. No. 08/643,804 filed 30 May 6, 1996 titled "Magnetic Spin Injected Field Effect Transistor and Method of Operation," (now U.S. Pat. No. 5,654,566).
The above applications and materials are expressly incorporated by reference herein. 35
The government of the United States may have certain limited rights in the present invention.
FIELD OF THE INVENTION
40
The invention relates generally to magneto-electronic devices which can be used for a variety of applications, including as memory elements in a random access memory array. In particular, the present invention is directed to a hybrid Hall Effect device which includes a ferromagnetic 45 material magnetically coupled to a conventional Hall Effect plate. The hybrid device can be coupled directly to a bias source and combined with other semiconductor structures, such as transistor isolation elements. The resulting array architectures have improved cross-talk, signal level and 50 signal-to-noise performance characteristics.
BACKGROUND OF THE INVENTION
Hybrid Hall Effect devices are described in my earlier 55 application (4) above. The present invention is directed to specific embodiments of such devices, as well as preferred arrangements of the same which result in improved memory elements and performance of memory arrays.
By modifying a classic Hall plate, and in particular by 60 coupling it to a ferromagnetic material, novel hybrid devices can be created that are useful for applications in high density nonvolatile memory and logic gate environments. A schematic figure that can be used to represent the novel Hybrid Hall Effect Device 10 is illustrated in FIG. 1. A standard 65 representation for a classic four-terminal Hall plate 12 is a cross centered in a square. Two opposing terminals are used
2
for current bias (or voltage bias), for example terminals 14 and 16, and the other two terminals 18 and 20 are used for sensing a bipolar Hall voltage.
One embodiment of the hybrid Hall Effect device described in my earlier application (4) incorporates a ferromagnetic film F 22 fabricated to be electrically isolated from the Hall plate but covering a portion of the area of the Hall plate such that an edge 26 of the film is over a central region of the plate. Local, fringe magnetic fields from the edge of the ferromagnetic film are perpendicular to the plane of the plate, may point "up" or "down" depending on the orientation of the magnetization in F, and have an average value Bav in the active region of the device. For constant bias current the sensed Hall voltage has opposite polarity when the fringe fields are "up" compared with when they are "down." The magnetization "M 24 of F is typically in the plane of F and lies along an axis parallel with that of the bias current. Other orientations can be used however, such as magnetizations that are perpendicular to the plane instead. The magnetization can be configured to have two stable states along this axis, with the two states corresponding to "up" or "down" fringe fields near the edge of F, positive or negative Hall voltage, and thus representing a binary bit of information "1" or "0". The magnetization state can be set (written) to be positive or negative by using the magnetic field associated with a positive or negative current pulse transmitted down an integrated write wire that is contiguous with F, discussed in detail in application (4) above, as well as below. It follows that such a device can be used as the nonvolatile storage element in an array of elements comprising a nonvolatile random access memory (NRAM). FIG. 1 depicts the first embodiment of the Hybrid Hall Effect device, generally referred to hereafter as a "modified hall plate." Again, while application (4) describes one preferred physical embodiment of the modified hall plate, it will be understood by those skilled in the art that a variety of layer materials, layer structures and layer arrangements are possible.
My earlier application (4) described various arrangements of the type of hybrid Hall Effect devices described above which could be used as a memory array. In particular, a linear row of elements can be constructed with such devices, so that the positive current bias terminal of one element (e.g. terminal 14 in FIG. 1) is connected in series with the negative current bias terminal of another (e.g. terminal 16). In this way, a number n,- of elements is combined in one row and biased by a single current (or voltage) supply source, \B or VB. Each element of that row can then be sensed by a unique sense amplifier devoted to that element. Alternatively, a single amplifier can be used for all elements in the row if a selection and isolation device is used to isolate each element from all of the other elements. A field effect transistor (FET) used in this manner can be referred to as a "select transistor."
While the above memory array arrangement is satisfactory for many environments, other applications may experience undesirable problems when a number Uj of rows are fabricated together to form an array of this type. In particular, because the hybrid Hall Effect memory cells are not isolated from each other, Hall voltages generated at one cell can dissipate through common connections to neighbor cells and the signal level, as well as the signal to noise ratio (SNR) of the readout voltage for every cell can be degraded. There is a need, therefore, for hybrid Hall Effect memory cell embodiments and arrangements which eliminate and/or reduce such performance problems.
3
SUMMARY OF THE INVENTION
An object of the present invention therefore is to provide an improved architecture for a hybrid Hall Effect device to be used as a memory cell in a memory array. 5
A further object of the present invention is to provide an improved architecture for a random access memory array of hybrid Hall Effect memory cells which permits a bias voltage (or current) to be provided to each such cell, thus resulting in improved signal level and signal-to-noise ratio 10 characteristics for such array.
Another object of the present invention is to provide an improved memory cell for application in a memory array environment, which affords improved cell isolation and reduced cross talk between cells in the array. 15
According to the present invention, an array of modified Hall Effect plate memory elements with improved performance is formed by providing the bias source \B or (VB) directly to each cell in the array to increase the SNR. The outputs of the Hall Effect plate memory elements can be 20 coupled to a select transistor which acts as an isolation element to reduce cross-talk between cells in the array. In this way, the magnetization state of the ferromagnetic layer and the output of the Hall Effect plate for each hybrid memory cell is retained and isolated from other hybrid 25 memory cells.
The bias source and ground for the cells are provided by additional individual levels of lithography, and this choice increases the number of fabrication steps (and levels) by comparison with the simpler array described earlier. How- 30 ever, the hybrid Hall Effect device requires far fewer levels than conventional dynamic random access memories (DRAM) and this economy of fabrication is only marginally diminished by the addition of a few levels.
In the preferred cell architecture, one selection transistor 35 is fabricated within each cell. When a select transistor for a particular cell is addressed the Hall voltage developed by that cell is the only voltage transmitted to the input line of a sense amplifier that is common to a column of cells. The signal to noise ratio (SNR) can be adjusted to any desired 40 value and there is minimal (or zero) signal leakage between neighboring cells.
Thus, the preferred cell architecture is comprised of a single modified hall plate and an accompanying single select FET that can be fabricated with a small cell area and a high 45 packing density. Furthermore, in another embodiment of the invention, if readout voltages are sufficiently large as to accommodate some dissipation as the voltage is transmitted across the FET, then select transistors with relatively poor operating parameters, or other high impedance elements 50 such as resistors or diodes can be used. Such transistors and high impedance elements can be made by thin film processing, and this would permit the fabrication of multiple layers of cells, further increasing density.
Data can be written to and stored in the hybrid Hall Effect 55 memory element by transmitting a suitable current pulse to alter the magnetization state of the ferromagnetic layer. In addition, write circuitry can be included in a memory array of the present invention if it is desirable that the data (magnetization state) be modifiable at a later time. The 60 stored data is non-volatile and will not be affected by loss of power to the memory cell array.
BRIEF DESCRIPTION OF THE DRAWINGS
65
FIG. 1 is a schematic representation of a typical hybrid Hall Effect device of the present invention;
4
FIG. 2 depicts one embodiment of the present invention consisting of an array of hybrid Hall Effect memory cells optimized for high SNR and minimal cross-talk.
FIG. 3 depicts another embodiment of the present invention consisting of an array of hybrid Hall Effect memory cells that has a higher packing density than that of FIG. 2 but which has lower SNR.
FIG. 4 illustrates a typical write circuit that may be employed in the present invention for writing data to the hybrid Hall Effect memory cells.
DETAILED DESCRIPTION OF THE
INVENTION
The hybrid Hall Effect device of the present invention consists basically of a ferromagnetic material coupled magnetically to a conventional Hall Effect plate. It should be noted that the term "Hall Effect plate" is a typical term in the art, and as used herein is intended in its broadest sense to include any device or structure that can be used to generate a Hall effect signal. For example, as noted above, Hall Effect plates typically include a single layer of semiconductor material and four associated terminals of some kind. It will be understood by those skilled in the art, however, that a variety of structures, materials and arrangements may be employed to create a device or structure that can generate a Hall Effect. For example, the conductive layer and terminals of the plate can be fabricated in a semiconductor material using a variety of conventional processing techniques. The teachings of the previous and present invention can be applied generally, therefore, to any such Hall Effect devices or structures irrespective of their form.
Accordingly, in the present invention as seen in FIG. 1, Hall Effect plate 12 includes a conductive layer that is capable of carrying a current when a bias voltage is applied to terminals 14 and 16. When the ferromagnetic material 22 is set to a particular magnetization state, this state is magnetically coupled to a portion of conductive layer 12, and results in a hall effect voltage between terminals 18 and 20. In a preferred embodiment, ferromagnetic film 22 overlies approximately ¥i of the area of plate 12, and a magnetic field emanating from an edge portion of such film is located and coupled substantially perpendicular to an axis extending from terminals 18 and 20. If terminal 18 is set to a ground reference, it can be seen that a signal is output at terminal 20 by the Hall Effect plate that is representative of the magnetization state of the ferromagnetic material. This magnetization state can have different values corresponding to different values for a data item to be stored in the memory element.
In one embodiment of the present invention the magnetization state of the ferromagnetic material is configurable and can be set and reset with an accompanying write circuit shown generally in FIG. 4. There may be memory applications, however, where it is only necessary to set the magnetization state once to store a particular data value, such as in a ROM, and in such cases, a write circuit is not necessary. It can be further observed that once the magnetization state of the ferromagnetic material 22 is set, it is non-volatile and thus stores the value for such data item permanently at such memory location until such time as it is desired to retrieve and read such item.
An improved array of hybrid Hall Effect devices that can be used as memory cells in an NRAM is depicted schematically in FIG. 2 and permits an analysis of the process by which data can be read from such cells. For simplicity of discussion a 3x2 array is presented but those skilled in the
5
art will appreciate that the principles for an array of arbitrary size n rowsxm columns are exactly the same. A bias voltage is supplied directly to each cell. In memory cell 100, for example, the bias voltage is applied across terminals 110 and 112. It is understood by those skilled in the art that the 5 supply voltage may be bipolar (V+ to V- as drawn) or may have a single polarity and a ground. The reference ground for sensing the Hall voltage generated by the memory cell is attached to a third terminal in each cell, e.g. terminal 1.14 in cell 100. Again, it is understood by those skilled in the art 10 that the reference ground may be, in some applications, electrically equivalent with the bias voltage ground.
The fourth terminal of each memory cell, e.g. terminal 116, provides the Hall voltage as a readout signal of magnitude Vr. Again, it will be understood by those skilled in the 15 art that this hall signal output "terminal" 116 (as with the other terminals 110, 112 and 114 of the Hall Effect device) may take on a variety of forms, and be fabricated and coupled to the modified Hall plate in a variety of ways using conventional semiconductor processing techniques. 20
The output signal of terminal 116 is connected to one terminal of a select transistor 120, which, in a preferred embodiment, is fabricated as part of memory cell 100 using conventional semiconductor processing techniques to increase the packing density of the array. It is further well 25 known in the art that the output signal (readout voltage) from the Hall Effect device in memory cell 100 may be bipolar or, if a suitable geometric offset is lithographically employed during fabrication, such readout may have two values such as zero and positive or zero and negative. The other terminal 30 of select transistor 120 is connected by line 122 to the input of a sense amplifier 124. In the array of FIG. 2, select transistor 120 (and all transistors that perform a selection (or address) function) are depicted as n-channel enhancement mode FETs. When an adequate voltage is applied to the gate 35 of such a device the channel between source and drain has high conductance and acts as a short circuit. With zero voltage applied to the gate the channel has high impedance and the source and drain terminals are electrically isolated by an open circuit. It is understood to those skilled in the art 40 that a variety of transistor devices or similar switching devices could be used to accomplish this same function depending on the requirements of the particular application.
The array is provided with conventional row select 130 and column select 132 logic such that any individual cell (i, 45 j) can be uniquely addressed, and the data read out from the hybrid Hall Effect memory cell. For example, to address cell 100 shown in the first row of the array, the gates of all select FETs in such row are raised to a level VDD high enough to activate the FETs. When row select transistor 140 is thus 50 activated, a reference ground for terminal 114 (and for all cells in the first row) is connected to a global reference ground 146. Similarly, when row select transistor 142 is activated, terminal 110 of cell 100 (and all such terminals for all cells in the first row) receives a positive bias through the 55 connection to a common positive voltage supply 148 V+. When row select transistor 144 is activated, the negative bias terminal 112 for cell 100 (and for all cells in the first row) is connected to a common negative voltage supply 150 V-.
In some memory applications where high speed access of 60 the data is more desirable than low power, it may be desirable to maintain (or fix) the bias voltage across terminals 110 and 112 at all times, or to pre-connect the bias voltage prior to reading the cell using well-known techniques in the art. In such embodiments where a bias is 65 already applied to cell 100, the state of the memory cell is represented by two separate measurable quantities, includ
6
ing both the magnetization state of the ferromagnetic material and the associated generated Hall Effect signal at terminal 116. The electrical signal representing the value of the data stored in the memory cell in these embodiments is present and can be sensed immediately, thus increasing the speed of operation of memory arrays constructed in this fashion.
Next, a row selection signal is applied to select transistor 120 of cell 100 (and to each select transistor for the cells in the first row). This signal activates select transistor 120 and a Hall voltage generated by the modified Hall plate of each cell is transmitted to the sense amplifier for the column which includes that particular cell. For cell 100 therefore select transistor 120 is activated and the readout voltage is transmitted to sense amplifier 124. Sense amplifier is a conventional circuit, and typically amplifies the readout voltage Vr to a level appropriate for related logical circuitry (CMOS or TTL). The address and readout process is completed by choosing a particular column. To read out the contents of cell 100, gate voltages of select transistors at the third column of the array are raised to a level VDD adequate to activate the FETs. Activating output FET 154, for example, transmits the amplified readout voltage to a readout terminal 160 at an appropriate CMOS or TTL level such that it can be incorporated into other logical or processing operations. This is but one example of a typical sense amplifier and read out circuit that could be used with the present invention. A number of well-known address logic, sense amplifier and readout circuits can be used with the present invention depending on the requirements of a given application.
It is clear from this example that an array of this kind permits a high degree of cell isolation and a high value of SNR. The supply voltage, bias ground and reference ground of each cell in a row are fixed to common values during readout, and are isolated from all other rows.
The output readout voltage Vr output of each cell is the only floating voltage in the cell and Vr is isolated from other cells in the row. Furthermore, it is isolated from other cells in its column by the select transistor 120, which acts as an isolation element. Thus, there is no "cross-talk" between cells.
Whereas the cell utilized in the array of FIG. 2 requires a single select transistor and is comprised of only two elements, thereby permitting fabrication with a minimal area and promoting high packing density, there may be applications wherein a further reduction of cell area is desired. The select transistor within each cell can be eliminated with some degradation of performance. The array depicted in FIG. 3 is able to minimize the "cross-talk" between neighboring cells even when there is no select transistor for such cell. In most ways the array of FIG. 3 is identical with the array of FIG. 2, but the fourth terminal of each cell, e.g. terminal 212 in cell 200 outputs a Hall voltage as a readout signal of magnitude Vr and is connected through a high impedance element (such as a thin film resistor or diode) to a sense amplifier 214 used by all cells of that column.
The address and readout operations for this embodiment proceed in a manner similar to the discussion above. In contrast with the almost complete isolation afforded by the array of FIG. 2, a fraction of the voltage Vr of cell 200 is transmitted by terminal 220 to cell 250 and then by terminals 222,232,224,234,226 and 236 to cell 260 and by terminals 222, 242, 224, 244, 226, and 246 to cell 270, and finally by terminals 230 and 252 to cell 280 and by terminals 240 and 254 to cell 290. In this way the Hall voltage generated in each cell interferes somewhat with that of its neighboring
|
