Non-volatile, piezoelectronic memory based on piezoresistive strain produced by piezoelectric remanence

Abstract

A nonvolatile memory storage device includes a ferroelectric (FE) material coupled with a piezoresistive (PR) material through an inherent piezoelectric response of the FE material, wherein an electrical resistance of the PR material is dependent on a compressive stress applied thereto, the compressive stress caused by a remanent strain of the FE material resulting from a polarization of the FE material, such that a polarized state of the FE material results in a first resistance value of the PR material, and a depolarized state of the FE material results in a second resistance value of the PR material.

Description

BACKGROUND

The present invention relates generally to integrated circuit devices and, more particularly, to a non-volatile, piezoelectronic memory based on piezoresistive strain produced by piezoelectric remanence.

Complementary Field Effect Transistors (FETs) support the standard computer architecture (CMOS) currently used in logic and memory. FETs exploit high channel mobility to control few-carrier currents electrostatically. However, limitations in this highly successful technology are appearing at current and future device scales. In particular, the inability to operate with power supplies significantly below 1 volt (V) limits device speed because faster clock speeds imply unacceptably high power consumption. Thus, it would be highly desirable to develop new switches and memories enabling computer architectures operable at low voltages/powers and high speeds.

SUMMARY

In an exemplary embodiment, a nonvolatile memory storage device, includes a ferroelectric (FE) material coupled with a piezoresistive (PR) material through an inherent piezoelectric response of the FE material, wherein an electrical resistance of the PR material is dependent on a compressive stress applied thereto, the compressive stress caused by a remanent strain of the FE material resulting from a polarization of the FE material, such that a polarized state of the FE material results in a first resistance value of the PR material, and a depolarized state of the FE material results in a second resistance value of the PR material.

In another embodiment, a nonvolatile memory cell includes a storage transistor coupled to an access transistor, the storage transistor comprising a first ferroelectric (FE) material coupled with a first piezoresistive (PR) material through an inherent piezoelectric response of the first FE material, wherein an electrical resistance of the first PR material is dependent on a compressive stress applied thereto, the compressive stress caused by a remanent strain of the first FE material resulting from a polarization of the first FE material, such that a polarized state of the first FE material results in a first resistance value of the first PR material, and a depolarized state of the FE material results in a second resistance value of the first PR material that is higher than the first resistance value, wherein the first FE material is disposed between first and second electrodes, and the first PR material is disposed between the second electrode and a third electrode; wherein the first FE material is polarized by initial application of a voltage across the first FE material to result in an initial polarization D_s, and thereafter by removal of the voltage to leave the first FE element with a remanent polarization D_r and a remanent strain S_r; and wherein the first FE material is depolarized and the remanent strain removed by applying an alternating voltage of decreasing amplitude across the first FE material.

In another embodiment, a multibit, nonvolatile memory storage device includes a first ferroelectric (FE) material and a second FE material coupled with a piezoresistive (PR) material through an inherent piezoelectric response of the first FE material and second FE material, and arranged in a single stack, wherein an electrical resistance of the PR material is dependent upon on a compressive stress applied thereto, the compressive stress caused by a remanent strain of the first FE material and the second FE material resulting from a polarization of the first FE material and a polarization of the second FE material such that the PR material is set to assume one of a first resistance value, a second resistance value, a third resistance value of the PR material, and a fourth resistance value.

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

Referring to the exemplary drawings wherein like elements are numbered alike in the several Figures:

FIG. 1 is a perspective view of an exemplary memory storage device in accordance with an exemplary embodiment;

FIG. 2A is a graph illustrating hysteretic polarization of the ferroelectric (FE) element of the memory storage device of FIG. 1;

FIG. 2B is a graph illustrating depolarization of the FE element of the memory storage device of FIG. 1;

FIG. 2C is a graph illustrating polarization and depolarization of the FE element of FIG. 1 in an imprinted state, according to another embodiment;

FIG. 3A is a graph illustrating polarization p vs. electric field E for a FE material, PMN-33% PT;

FIG. 3B is a graph that illustrates the strain of the FE material transverse to its thickness versus electric field;

FIG. 4 is a graph that illustrates the resistive response to pressure in a SmSe piezoresistive (PR) medium;

FIG. 5 is a schematic diagram of a two transistor memory cell that may be used in a cross-point memory matrix to read and write a bit using the storage element design of FIG. 1, in accordance with an exemplary embodiment;

FIG. 6 is a table that illustrates read and write operations for the memory element of FIG. 5;

FIG. 7 is a schematic diagram of a memory matrix utilizing the type of memory elements shown in FIG. 5;

FIG. 8 is a schematic diagram of a cross-point FePZRAM matrix that utilizes a single FE-PR stack per memory element, in accordance with another embodiment;

FIG. 9 is a pressure-resistance curve for a Ca₂RuO₄ PR material;

FIG. 10 is a schematic, cross sectional view of a multibit memory storage element, in accordance with a further embodiment;

FIG. 11 is a schematic diagram of an exemplary cross-point array for the multibit memory element of FIG. 10;

FIG. 12 is a schematic, perspective view of another multibit memory storage element, in accordance with a further embodiment;

FIG. 13A is a graph illustrating hysteretic polarization of a first piezoelectric material of the memory storage device of FIG. 12;

FIG. 13B is a graph illustrating hysteretic polarization of a second piezoelectric material of the memory storage device of FIG. 12;

FIG. 14 is a pressure versus resistance curve illustrating four separate programmable states for the memory storage device of FIG. 12; and

FIG. 15 is a graph comparing programming time versus memory density for various memory technologies.

DETAILED DESCRIPTION

Disclosed herein are embodiments of a non-volatile, ferroelectric memory having a piezoresistive read capability (FePZRAM) which makes the device compatible with piezoelectronic, low-voltage switching technology. Referring initially to FIG. 1, there is shown a perspective view of an exemplary memory storage device 100 in accordance with an exemplary embodiment. The memory storage device 100 includes a ferroelectric (FE) element 102 such as PZT (Pb[Zr_xTi_1-x]O₃) coupled with a piezoresistive (PR) element 104, such as SmSe. FE materials are a subset of piezoelectric (PE) materials which, after being driven to a saturation voltage by an external source, retain a remanent polarization after removal of the external voltage source. Thereby, a remnant strain state, with respect to the unpolarized state, is also achieved within the FE material.

The FE element 102 is disposed between a gate contact (electrode) 106 and a common contact 108, and the PR element 104 is disposed between the common contact 108 and a sense contact 110. A high yield strength medium serves as a mechanical clamp 112 around the memory structure. As further illustrated in FIG. 1, the FE element 102 has a cross sectional area (A) along the x-y axes and a thickness (L) along the z-axis. The PR element 104 has a cross sectional area (a) along the x-y axes and a thickness (l) along the z-axis.

The ferroelectric element 102 is operated in two states. First, in a “Bit 1” state, the FE element 102 is polarized in a direction parallel to the z-axis, and thus is elongated along the same axis. In a “Bit 0” state, the FE element 102 is unpolarized, and thus not elongated. The FE 102 is polarized (as shown in the graph of FIG. 2A) by applying a suitable voltage across the gate-common pair of contacts 106/108 to initially bring the polarization to D_s, and thereafter then turning the applied voltage off to leave the FE element 102 with both a remanent polarization D_r and a remnant strain S_r. As shown in the graph of FIG. 2B, the FE element 102 may be depolarized by applying an alternating voltage of steadily decreasing amplitude.

In an alternative embodiment, the FE element 102 may be placed in an imprinted state, as shown in the graph of FIG. 2C, with two positive remanent polarizations D_r1 and D_r2. To go from polarization D_r2 to D_r1, an electric field is applied to bring the polarization to D_s and then the electric field is reduced to zero leaving the polarization at D_r1. To go from polarization D_r1 to D_r2, a negative electric field is applied to bring the polarization to −D_s and then the field is increased to zero leaving the polarization at D_r2. It should be noted that the embodiments described below are specific to the non-imprinted case in which polarization is illustrated as in FIG. 2A and depolarization is illustrated as in FIG. 2B.

An exemplary switching time for the FE element 102 may be on the order of 10's of picoseconds (ps) for a thin FE film and voltages on the order of several tens of millivolts (mV). Switching occurs by the motion of a domain wall at the speed v=2.4×10⁻⁴ E·m/s for an electric field E in V/m. For V=0.05 V and thickness L=30 nm (FIG. 1), this speed is 400 m/s and the propagation time is 75 ps.

FIG. 3A illustrates a plot of polarization p vs. electric field E for a FE material, PMN-33% PT, which is also a piezoelectric material (an essential additional property utilized in the present embodiments). In FIG. 3B, the graph illustrates the strain transverse to its thickness versus electric field. It is seen that by applying a moderate electric field, e.g., 7 kV/cm (equivalent to 0.07 V across a 100 nm thick film), the film is electrically polarized and a strain S in the film transverse to its thickness of 0.003 can be produced on relaxing back to zero field. This zero-field strain is used to store a bit.

When polarized, the expansive strain acting on the FE element 102 is used to compress the PR element 104, which is in contact with the FE element 102, and constrained by the surrounding rigid clamp 112 to cause compression rather than expansion of the whole structure. The compression of the PR element 104 due to the “Bit 1” polarization of the FE element 102 is used to provide a piezoresistive read signal through the sense and common contacts 110, 108.

Referring now to FIG. 4, there is shown a graph that illustrates the resistive response to pressure in a proposed PR medium, SmSe, which undergoes a large change in resistivity under an applied GPa-magnitude pressure. Other materials which may be suitable in this regard include, for example, SmS, SmTe, TmTe, CaRuO₄, Ni(S_1-xSe_x)₂, and V₂O₃ doped with 4% Cr.

An approximate expression for the stress in the PR due to the polarization-induced strain in the FE is:

$T_{\mathrm{PR}}=\frac{S\left(p\right)}{\frac{l}{{\mathrm{LY}}_{\mathrm{PR}}}+\frac{a}{{\mathrm{AY}}_{\mathrm{FE}}}}$

Here, T_PR is the stress in the PR, S(p) is the remanent strain in the FE due to polarization p, the dimensions and areas (l, L, a and A) are as described in FIG. 1, Y_PR and Y_FE are the Young's moduli of the PR element 104 and FE element 102, respectively. As an example, if it is assumed that the thickness ratio l/L is 1/10, and the area ratio a/A is 1/16, and Young's moduli are Y_PR=100 GPa, Y_FE=20 GPa, then using the FIG. 3B example with S(p)=0.003,T_PR=0.727 GPa.

Referring to the exemplary PR response in FIG. 4, it is seen that a pressure of 0.7 GPa will produce a change in resistivity of approximately log₁₀ r_ON/OFF=1.7 orders of magnitude, which is more than sufficient to present a clear read signal.

Referring now to FIG. 5, there is shown a schematic diagram of an individual, two transistor memory cell 500 that may be used in a cross-point memory matrix to read and write a bit using the storage element design of FIG. 1. The memory cell 500 includes a pair of piezoelectronic transistors (PETs), T₁ and T₂. Both PETs are surrounded either individually or together by a High Young's—modulus Material (HYM) (not specifically shown in FIG. 5). Transistor T₁ serves as the storage element and corresponds to the device 100 illustrated in FIG. 1, while transistor T₂ serves as an access transistor for T₁, and in an exemplary embodiment is implemented as a four-terminal PET.

In brief, the 4-terminal PET T₂ includes a piezoelectric (PE) crystal material in lieu of the FE material of the storage transistor T₁. In addition, instead of being disposed between a gate electrode and a common electrode, the PE material of T₂ is disposed between a pair of gate electrodes. A low permittivity insulator layer (not shown) separates the one gate electrode from a first sense electrode. Effectively, the common electrode of the 3-terminal configuration is split into a second gate electrode and second sense electrode for a 4-terminal configuration. Additional details concerning 4-terminal PET devices may be found in U.S. application Ser. No. 13/719,965, the contents of which are incorporated herein by reference in their entirety.

The memory element 500 is coupled to a pair of control lines, b (bit line) 502 and w (word line) 504. The gate electrode (g₁) and the sense electrode (s₁) of the ferroelectric PET (FPET) storage transistor T₁ are coupled to ground. The common electrode (c₁) of the FPET storage transistor T₁ is coupled to a first sense electrode (s_2a) of the 4PET access transistor T₂, while a second sense electrode (s_2b) of the 4PET access transistor T₂ is coupled to the bit line 502. A first gate electrode (g_2a) of the 4PET access transistor T₂ is coupled to the word line 504, and a second gate electrode (g_2b) of the 4PET access transistor T₂ is coupled to ground. A switch 506 may be included in the peripheral logic to couple to separate read (sense) and write voltages (V_s, V_w) for the bit line b. The switch 506 may also be implemented using a PET device in lieu of an FET device. Internally, the gate electrode (g₁) of FPET T₁ is connected to one side of the FE capacitor, with the other side of the capacitor representing the common electrode (c₁). The resistor symbol of FPET T₁ represents the PR material connected between the common and sense contacts.

FIG. 6 is a table that illustrates read and write operations for the memory element 500 of FIG. 5. In a standby state, both the bit line 502 (b) and the word line 504 (w) are grounded (i.e., set to 0 volts). For a read operation, the bit line 502 is connected via a current sensor (I) to a small sense voltage source V_s=ε. A voltage H is applied to the word line, the effect of which is to turn T₂ ON (i.e., the access transistor T₂ couples the storage transistor T₁ to the bit line 502). The current through the sensor (I) detects the bit state of T₁; specifically, a large current indicates that the PR of T₁ is in the compressed low resistance state characteristic of a bit 1 value, while a low current indicates that the PR of T₁ is in the uncompressed high resistance state of a bit 0 value.

In a write mode of operation, the switch 506 uncouples the bit line 502 from the small sense voltage source V_s and couples the bit line 502 to a write voltage source V_w. To write a 1 to the storage transistor T₁, the voltage on the word line 504 is raised to H to turn T₂ ON. A constant write voltage H is also applied to the bit line 502 to polarize the FE material of the storage transistor T₁, and is thereafter turned off. To write a 0 to the storage transistor T₁, the applied write voltage on the bit line 502 is an oscillatory voltage (symbolized by ˜ in FIG. 6) with a gradually decreasing amplitude. It is not necessary that the FE be completely depolarized to represent bit 0, so long as the low polarization state results in a PR resistance of the storage transistor T₁ being distinguishable from the high polarization state, which again is facilitated by the exponential dependence of PR resistance on pressure (FIG. 4).

FIG. 6 also illustrates a “no write” state, in which the bit line 502 may be used for writing another memory cell (not shown) on the bit line 504 but associated with a different word line. So long as the voltage on the word line 504 is 0, there is no access to the storage transistor T₁ since the access transistor T₂ remains in an OFF state.

Memory matrix operation defines conditions on PET design. To detect the ON state, the condition R_2,ON<<R_1,ON<<R_1,OFF would be desirable. To detect a 1 bit in an enabled row, which is a row with T₂ in the low-resistance state, R_2on, the resistance in T₁, which is R_1,on, must be less than the parallel sum of all the off resistances in the T₂ 4 PETS that are in the other rows. In this way, the current through the bit to be read dominates the combined leakage current from the other rows. If N is the number of rows, then this readability condition is R_1,on<<R_2,off/N.

Leakage current through the piezoelectric, which is nominally a capacitor, should be minimized to avoid excessive energy dissipation. The leakage current through a PMN-PT piezoelectric with an electric field of 4×10⁶ V/m equals 10⁻⁵ Å/cm². This value is also expected for a design where the applied voltage is about 0.1 V and the PE thickness is about 25 nm. With a leakage current of 10⁻⁵ Å/cm² for each PET, a memory with 10⁹ PETs, each with an area of (30 nm)², has a total leakage current of 10⁻⁷ Å, or a total leakage power of 10⁻⁸ Watts. This is negligible compared to the other power dissipated.

FIG. 7 is a schematic diagram of a memory matrix 700 utilizing the type of memory elements 500 shown in FIG. 5. Assuming the memory matrix 700 of FIG. 7 is M×N, with M˜N˜1000, and a lithography scale of 10 nm, the amount of energy used to charge the M bit lines (b(1), b(2), etc.) is larger than the amount of energy used to charge a word line (w(1), w(2), etc.). More specifically, the energy is dominated by the wiring capacitance and is on the order of about 1×10⁻¹³ J for 1000 bit lines. This also gives a negligible contribution of 1×10⁻⁴ Watts to the power at an input bit rate of 1 Gb/s per bit line. For sources of equal resistance, the charging time of the word line is longer than that of the bit line, which is dominated by the input capacitances of the 4-PETs, and is on the order of about 1×10⁻¹⁶MR_w where R_w is the word line source impedance.

Referring now to FIG. 8, there is shown a schematic diagram of a cross-point FePZRAM matrix 800, in accordance with another embodiment. This embodiment utilizes just a single FE-PR stack per memory element 802 (and surrounded by a HYM, not shown), which allows for approximately 4 times the density on a chip with respect to the memory element 500 illustrated in FIGS. 5 and 7. Each memory element 802 in FIG. 8 includes a bottom ferroelectric layer (Feb) with opposing electrical contacts on each side, in mechanical but insulated contact with a top ferroelectric layer (FEt) that also has two separate electrical contacts on opposing sides thereof. In addition, a top PR layer is in mechanical contact with the top ferroelectric layer (FEt), wherein the top PR layer also has electrical contacts on opposing sides thereof, one of which may be shared with the adjacent electrical contact of FEt. There are two word lines, wt(n) and wb(n) (for n equal to rows 1, 2, 3, etc.), providing electrical contact to each row in the matrix 800, as well as one bit line, b(m) (for m equal to columns 1, 2, 3, etc.) providing electrical contact to each column. Each word line wb(n) connects to the top electrical contact on FEb, the other contact of which is grounded, and each word line wt(n) connects to the bottom electrical contact of FEt and to the top electrical contact of the PR. Each bit line b(m) contacts the conducting layer between the associated FEt and PR.

Each memory element 802 shown in FIG. 8 utilizes a type of PR that has a slowly declining resistance with increasing pressure at low pressure, and then a steeply declining resistance with increasing pressure at high pressure, as shown in the pressure-resistance curve of FIG. 9. Specifically, FIG. 9 is a pressure-resistance curve for the material Ca₂RuO₄. As is illustrated, a half compression of the PR reduces its resistance by only a small amount, while a full compression reduces it by a large amount.

Referring again to FIG. 8, operation of the memory elements 802 may be understood as follows. Each bottom FEb is controlled by the associated bottom word line wb(n) and is placed in a non-polarized state for all write operations and for all read operations in any other row but its own. However, bottom word line wb(n) places FEb in a polarized state when the n^th row is being read. This operation of polarizing and unpolarizing FEb takes the place of the row-enabling switch (T₂) in the memory element of FIG. 5. The top word lines wt(n), in conjunction with the bit lines b(m), control the value of the bit in the (m,n) memory unit by setting FTt(m,n) to a polarized or unpolarized state. This write operation can be done simultaneously for all memory elements in the row.

The read operation is implemented by first setting all of the bottom ferroelectrics in a row, FEb, to a polarized state, thereby compressing the PR halfway. Each FEb in the non-read rows remains unpolarized. A small voltage c is then applied to the top word line wt(n) for row n, and the bit lines are set to zero voltage with a current sensor. Current runs through each PR in the row to be read and then to the sensors in the bit lines, but significant current does not run through the other rows because their resistances are high. A top ferroelectric FEt that is polarized compresses the PR the other half of the way, dropping its resistance significantly so that a large current is read by that bit line sensor. A top ferroelectric FEt that is unpolarized does not compress the PR any more than the bottom FEb already compresses it, and therefore the PR resistance remains high. The current through that PR is therefore low in the bit line sensor. In this way, the states of polarization in all of the FEts in a row can be simultaneously determined by the sensors. This read operation is non-destructive because FEt is unchanged by the small applied voltage ε.

The standby state for all of the word lines and bit lines is zero voltage. The top ferroelectric, FEt, keeps its polarization state with zero voltage, and thus the memory is non-volatile. It should also be noted that the bottom FEb in each element could be replaced by a piezoelectric that requires a continuously applied voltage during the read-row step to compress the PR.

Referring now to FIG. 10, there is shown a schematic, cross sectional view of a multibit memory storage element 1000, in accordance with a further embodiment. In this embodiment, the element 1000 is capable of storing two bits. The memory storage element 1000 includes two FE layers, FE1 and FE2, wherein FE1 has a thickness of about twice that of FE2. Both FE1 and FE2 are controlled by a pair of gate contacts, and are insulated from one another. The electric field (and thus polarization and strain) in each FE element is selected to be the same, wherein twice the magnitude of voltage is applied across FE1 to reach saturation polarization with respect to FE2. The states of the system are depicted in the table below:

TABLE 1
States of 2 bit memory
polarization	none	2 (FE2)	1 (FE1)	1 + 2 (FE1 and FE2)
Δl	0	Δl0	2 Δl0	3 Δl0

In Table 1, Δl is the intrinsic length expansion of the FE1+FE2 stack under different polarization combinations, with Δl₀ being the expansion for FE2. Given the exponential behavior of resistivity vs. stress/strain in the PR characteristic graph of FIG. 4, these values may easily be made distinguishable in terms of read current. The 4 different states can therefore be read, which defines a 2-bit memory element. Expanding this concept further, three stacked FEs with thicknesses of L, 2L, and 4L, could store 8 states, thereby providing a Byte of data in a single device footprint. It should be readily apparent to those skilled in the art that other combinations of FE thicknesses and remnant polarizations, as well as other contact dispositions, are also possible to attain different numbers of states.

An exemplary cross-point array 1100 for the multibit memory element 1000 of FIG. 10 is shown in FIG. 11. For ease of illustration, the surrounding HYM is omitted for the elements 1000. The array 1100 of FIG. 11 is written in a similar manner to that of FIG. 8, except that there are now additional word and bit lines to control the additional FEs in each memory element 1000. A read operation for the array 1100 of FIG. 11 senses a small current from bit line b1(m) to bit line b2(m) when the bottom word line in the row, wb, is activated by a voltage. Other wiring arrangements are possible, however. As is the case with the FIG. 8 embodiment, the bottom FE layer in each element 1000 may be replaced by a piezoelectric layer.

Referring now to FIG. 12, there is shown a schematic, perspective view of another multibit memory storage element 1200, in accordance with a further embodiment. Similar to the embodiment of FIG. 10, the storage element 1200 includes a pair of FE layers, 1202-1 and 1202-2. A PR element 1204 is also in mechanical contact with the FE stack; however, unlike the FIG. 10 embodiment in which the FE layers are each controlled by a separate pair of electrodes and separated from one another by an insulator layer, FE layers 1202-1 and 1202-2 are in contact with one another and both controlled by a single pair of electrodes 1206, 1208 (similar to the FIG. 1 embodiment). The PR element is disposed between electrodes 1208 and 1210. Exemplary dimensions for the storage element in FIG. 12 may be X≈45-90 nm; Y≈45-90 nm; Z≈35-90 nm; and T≈6-12 nm.

In order to produce four different logic states in a nonvolatile manner, the piezo materials for FE layers 1202-1 and 1202-2 are selected to be different from one another. In addition, and as illustrated collectively by FIGS. 13A, 13B and 14, the dimensions and materials of the multibit memory storage element 1200 are tailored such that for a first logic state (level 1) both FE layers 1202-1 and 1202-2 are depolarized, resulting in a high resistivity cell. For a second logic state (level 2), the applied electric field is ≈E_S2 such that the remanent polarization is set to D_r2 and results in a displacement such that P₂<6.2 kbar. For a third logic state (level 3), the applied electric field is ≈E_S3 such that the remanent polarization is set to D_r3 and results in a displacement such that P₃<6.2 kbar, but still greater than P₂ in order to have a lower resistivity at level 3. Then, for the fourth logic state (level 4), the applied electric field is greater than E_S3 such that the remanent polarization is set to D_r4 and results in a displacement such that P₄>6.2 kbar, and resulting in a dramatic drop in resistivity.

FIG. 15 is a graph comparing programming time versus memory density for various memory technologies. Generally, these properties tend to anticorrelate, covering a range from the highest storage density/slowest speed NAND FLASH to lower density, faster technologies such as SRAM and MRAM. It is estimated that the presently disclosed technology, termed FePZRAM herein, will lie somewhere between DRAM and SRAM in terms of memory storage density and speed. However, this determination will ultimately be made upon further research. Advantageously, FePZRAM is a nonvolatile memory technology, unlike SRAM and DRAM. Another important consideration not illustrated in the graph of FIG. 15 is power, and here the distinction of FePZRAM is the capability to operate at low voltage and compatibility with the low voltage Piezoelectric technology. FePZRAM also is capable of being developed vertically to increase storage density (e.g., FIGS. 8, 10, 12). These capabilities are at present unique and make FePZRAM desirable for future low power computing technology.

While the invention has been described with reference to an exemplary embodiment or embodiments, it will be understood by those skilled in the art that various changes may be made and equivalents may be substituted for elements thereof without departing from the scope of the invention. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the invention without departing from the essential scope thereof. Therefore, it is intended that the invention not be limited to the particular embodiment disclosed as the best mode contemplated for carrying out this invention, but that the invention will include all embodiments falling within the scope of the appended claims.

Claims

1. A nonvolatile memory storage device, comprising: a storage transistor comprising a ferroelectric (FE) material coupled with a piezoresistive (PR) material through an inherent piezoelectric response of the FE material, wherein an electrical resistance of the PR material is dependent on a compressive stress applied thereto, the compressive stress caused by a remanent strain of the FE material resulting from a polarization of the FE material, such that a polarized state of the FE material results in a first resistance value of the PR material, and a depolarized state of the FE material results in a second resistance value of the PR material; andan access transistor comprising a second FE material, wherein the access transistor is formed in a single stack with the storage transistor and the second FE material is mechanically coupled to and electrically insulated from the first FE material.

2. The nonvolatile memory storage device of claim 1, wherein the remanent strain of the FE material is achieved by initial application of a voltage across the FE material to result in an initial polarization Ds, and thereafter by removal of the voltage to leave the FE material with a remanent polarization Dr.

3. The nonvolatile memory storage device of claim 2, wherein the remanent strain of the FE material is removed by applying an alternating voltage of decreasing amplitude across the FE material.

4. The nonvolatile memory storage device of claim 3, further comprising a mechanical clamp surrounding the FE material and the PR material such that the remanent strain in the FE material results in the application of the compressive stress to the PR material, with the first resistance value of the PR material being lower than the second resistance value of the PR material.

5. The nonvolatile memory storage device of claim 1, wherein the FE material is disposed between first and second electrodes, and the PR material is disposed between the second electrode and a third electrode.

6. A nonvolatile memory cell, comprising: a storage transistor coupled to an access transistor, the storage transistor comprising a first ferroelectric (FE) material coupled with a first piezoresistive (PR) material through an inherent piezoelectric response of the first FE material, wherein an electrical resistance of the first PR material is dependent on a compressive stress applied thereto, the compressive stress caused by a remanent strain of the first FE material resulting from a polarization of the first FE material, such that a polarized state of the first FE material results in a first resistance value of the first PR material, and a depolarized state of the FE material results in a second resistance value of the first PR material that is higher than the first resistance value, wherein the first FE material is disposed between first and second electrodes, and the first PR material is disposed between the second electrode and a third electrode;wherein the first FE material is polarized by initial application of a voltage across the first FE material to result in an initial polarization Ds, and thereafter by removal of the voltage to leave the first FE element with a remanent polarization Dr and a remanent strain Sr;wherein the first FE material is depolarized and the remanent strain removed by applying an alternating voltage of decreasing amplitude across the first FE material, andwherein the access transistor is formed in a single stack with the storage transistor, the access transistor further comprising a second FE material mechanically coupled to and electrically insulated from the first FE material, the second FE material disposed between fourth and fifth electrodes.

7. The nonvolatile memory cell of claim 6, wherein the access transistor further comprises: a piezoelectric (PE) material coupled to a second PR material, wherein an electrical resistance of the second PR material is dependent upon an applied voltage across the PE material by way of an applied pressure to the PR material by the PE material; andthe access transistor being coupled to a word line and a bit line such that activation of the word line causes the access transistor to facilitate read and write operations for the storage transistor.

8. The nonvolatile memory cell of claim 7, wherein for the access transistor: the PE material is disposed between fourth and fifth electrodes;the second PR material is disposed between sixth and seventh electrodes; andthe fifth and sixth electrodes are insulated from one another.

9. The nonvolatile memory cell of claim 6 wherein the first PR material comprises Ca2RuO4.

10. The nonvolatile memory cell of claim 6 wherein the second FE material is depolarized for all read and write operations on the storage transistors in other word lines, and the second FE material is polarized for a read operation of the storage transistor on its own word line.

11. The nonvolatile memory cell of claim 10, wherein polarization of the second FE material without polarization of the first FE material results in a first compression of the PR material, and polarization of the second FE material with polarization of the first FE material results in a second compression of the PR material, such that the resistance of the PR material with the first compression is larger than the resistance of the PR material with the second compression.