阅读说明：本技术 硅的背侧上的磁性隧道结(mtj)集成 (Magnetic Tunnel Junction (MTJ) integration on the backside of silicon ) 是由 S·马尼帕特鲁尼 T·戈萨维 I·扬 D·尼科诺夫 于 2019-05-28 设计创作，主要内容包括：本发明公开了一种存储器件,其包括具有正侧和背侧的衬底,其中,第一导线在所述背侧上,并且第二导线位于所述正侧上。晶体管在所述正侧上处于所述第二导线与衬底之间。磁性隧道结(MTJ)在背侧上处于第一导线和衬底之间,其中,所述MTJ的一端通过所述衬底耦合至所述晶体管,并且所述MTJ的相对端连接至所述第一导线,并且其中,所述晶体管进一步连接至所述正侧上的第二导线。(A memory device includes a substrate having a front side and a back side, wherein first conductive lines are on the back side and second conductive lines are on the front side. A transistor is between the second conductive line and a substrate on the front side. A Magnetic Tunnel Junction (MTJ) is between a first conductive line and a substrate on a backside, wherein one end of the MTJ is coupled to the transistor through the substrate and an opposite end of the MTJ is connected to the first conductive line, and wherein the transistor is further connected to a second conductive line on the front side.)

1. A memory device, comprising:

a substrate having a front side and a back side, wherein a first conductive line is on the back side and a second conductive line is on the front side;

a transistor on the positive side and between the second conductive line and the substrate; and

a Magnetic Tunnel Junction (MTJ) on the back side and between the first conductive line and the substrate, wherein one end of the MTJ is coupled to the transistor through the substrate and an opposite end of the MTJ is connected to the first conductive line, and wherein the transistor is further connected to the second conductive line on the front side.

2. The memory device of claim 1, wherein the MTJ is connected to the drain of the transistor and the source of the transistor is coupled to the second conductive line.

3. The memory device of claim 1 or 2, wherein the MTJ is connected to the drain of the transistor using a through via extending through the substrate from the front side to the back side.

4. The memory device of claim 3, wherein a gate of the transistor is coupled to a word line.

5. The memory device of claim 1 or 2, wherein the first conductive line comprises a bit line and the second conductive line comprises a source line.

6. The memory device of claim 5, wherein the bit lines include a read bit line and a write bit line.

7. The memory device of claim 1 or 2, wherein the memory device comprises a 1T-1MTJ Magnetic Random Access Memory (MRAM).

8. The memory device of claim 1 or 2, wherein the MTJ device comprises an SOT electrode comprising GSHE material.

9. The memory device of claim 8, wherein the GSHE material includes at least one of β -tantalum (β -Ta), β -tungsten (β -W), Pt, Hf, Ir, Bi, and doped Cu.

10. The memory device of claim 8, wherein the MTJ device comprises a free magnetic layer coupled with the SOT electrode, wherein the SOT electrode is coupled to a write bit line; and an opposite end of the MTJ device is coupled to a read bit line.

11. The memory device of claim 10, wherein the material stack comprising the MTJ device further comprises: a tunneling barrier, a fixed magnetic layer, a coupling layer, a Synthetic Antiferromagnet (SAF)/pinning layer, and a top electrode.

12. A memory device, comprising:

a substrate;

a backside of the substrate, comprising:

a read bit line and a write bit line; and

a Magnetic Tunnel Junction (MTJ) device; and

a front side of the substrate comprising:

a source line; and

a transistor controllable by a word line and coupled to the source line.

13. The memory device of claim 12, wherein the MTJ device on the back side comprises an SOT electrode comprising a spin hall effect material and a free magnetic layer in direct contact with the SOT electrode, wherein the SOT electrode defines one end of the MTJ device and is coupled to the write bit line; and a top electrode defines an opposite end of the MTJ device and is coupled to the read bit line.

14. The memory device of claim 12 or 13, wherein one of the drain/source terminals of the transistor on the front side is coupled to the SOT electrode on the back side through a via in the substrate.

15. The memory device of claim 14, wherein the punch-through via is connected to a via pedestal in contact with one end of the MTJ device.

16. The memory device of claim 12 or 13, wherein the memory device comprises a three terminal device, wherein the read bit line and the write bit line on the backside form a first terminal and a second terminal, and the source line forms a third terminal.

17. The memory device of claim 12 or 13, wherein one of the drain/source terminals of the transistor on the front side is coupled to the SOT electrode on the back side through the substrate and the other of the source/drain terminals is coupled to the source line on the front side.

18. The memory device of claim 12 or 13, wherein the word line is coupled to a gate terminal of a transistor.

19. The memory device of claim 12 or 13, wherein the memory device comprises a 1T-1MTJ Magnetic Random Access Memory (MRAM).

20. The memory device of claim 19, wherein the MTJ device comprises an SOT electrode comprising GSHE material.

21. The memory device of claim 20, wherein the GSHE material includes at least one of β -tantalum (β -Ta), β -tungsten (β -W), Pt, Hf, Ir, Bi, and doped Cu.

22. The memory device of claim 20, wherein the MTJ device comprises a free magnetic layer coupled with the SOT electrode, wherein the SOT electrode is coupled to a write bit line; and an opposite end of the MTJ device is coupled to a read bit line.

23. The memory device of claim 22, wherein the material stack comprising the MTJ device further comprises: a tunneling barrier, a fixed magnetic layer, a coupling layer, a Synthetic Antiferromagnet (SAF)/pinning layer, and a top electrode.

24. A method of fabricating an integrated circuit device, the method comprising:

forming a first substrate and a second substrate;

performing front-end processing on the first substrate to form a transistor;

performing backend processing over the transistor to form a first contact and a conductive backend layer on the first substrate;

performing front-end processing on the second substrate to form a Magnetic Tunnel Junction (MTJ);

performing back end processing over the MTJ to form a second contact and a conductive back end layer on the second substrate; and

attaching the MTJ from the second substrate to the first substrate.

25. The method of claim 24, further comprising: coupling one of the drain/source terminals of the transistor on a front side of the first substrate to the MTJ on a back side of the first substrate using a through via in the first substrate.

Technical Field

Embodiments of the present disclosure relate to the field of integrated circuit structures, and in particular, to the field of Magnetic Tunnel Junction (MTJ) integration on the back side of silicon.

Background

Scaling of features in integrated circuits has been a driving force behind the ever-growing semiconductor industry for the past decades. Scaling to smaller and smaller features enables the maximum density of functional units to be achieved on the limited chip area of a semiconductor chip. For example, shrinking transistor size allows an increased number of memory devices to be incorporated onto a chip, thereby producing a product with improved functionality. However, the driving of more and more functions is not without problems. The necessity to optimize the performance of each device becomes increasingly important.

Non-volatile embedded memory (e.g., on-chip embedded memory with non-volatility) enables energy and computational efficiency. However, leading edge embedded memory options, such as spin transfer torque magnetoresistive random access memory (STT-MRAM), may suffer from high voltage and high current density issues during programming (writing) of the cell. The density limitation of STT-MRAM can be attributed to the large write switching current and select transistor requirements. In particular, conventional STT-MRAM has cell size limitations due to the need for the drive transistor to provide sufficient spin current. Furthermore, such memories are associated with the large write current (>100 μ Α) and voltage (>0.7V) requirements of conventional Magnetic Tunnel Junction (MTJ) based devices. Specifically, this manifests itself as i) high write error rates or low speed switching (over 20ns) in MRAM based Magnetic Tunnel Junctions (MTJ), and reliability issues due to tunnel currents in the magnetic tunnel junctions.

Thus, significant improvements are still needed in MTJ-based non-volatile memory arrays.

Drawings

FIG. 1 shows a two terminal 1T-1MTJ (magnetic tunnel junction) bit cell for STT-MRAM.

FIG. 2A illustrates an integrated circuit including a 1T-1MTJ MRAM bit cell having a MOBS according to one embodiment of the disclosure.

FIG. 2B illustrates a typical material stack for a 1T-1MTJ bit cell based on GSHE Spin Orbit Torque (SOT) switching according to one embodiment of the present disclosure.

Fig. 2C is a top view of the device of fig. 2B.

Fig. 2D is a sectional view of the SOT electrode showing directions of spin current and charge current determined by SOT in metal.

FIG. 3 illustrates a memory device including a 1T-1MTJ MRAM bit cell with a MOBS in more detail according to one embodiment of the disclosure.

FIG. 4 is a top view of a layout of a cross-sectional view of a 1T-1MTJ MRAM bit cell with a MOBS according to one embodiment of the present disclosure.

FIG. 5 is a cross-sectional view of the 1T-1MTJ MRAM bit cell with the MOBS along line cross-sectional line AA of FIG. 4.

FIG. 6 is a graph of write energy-delay conditions for a 1T-1MTJ MRAM bit cell compared to a conventional MTJ according to one embodiment.

FIG. 7 is a graph of reliable write times for a 1T-1MTJ MRAM bit cell with a MOBS and a conventional MTJ according to one embodiment.

FIG. 8 is a flow chart representing various operations in a method of fabricating a 1T-1MTJ memory device having a MOBS according to embodiments disclosed herein.

Fig. 9A and 9B illustrate a wafer composed of a semiconductor material and including one or more dies having Integrated Circuit (IC) structures formed on a surface of the wafer.

FIG. 10 is a cross-sectional side view of an Integrated Circuit (IC) device assembly that may include one or more embedded non-volatile memory structures having a 1T-1MTJ memory device with a MOBS.

FIG. 11 illustrates a computing device according to one implementation of the present disclosure.

Detailed Description

Embodiments of a filter layer for Magnetic Tunnel Junction (MTJ) integration on the back side of silicon are described. In the following description, numerous specific details are set forth, such as specific materials and processing schemes, in order to provide a thorough understanding of embodiments of the present disclosure. It will be apparent to one skilled in the art that embodiments of the present disclosure may be practiced without these specific details. In other instances, well-known features, such as single or dual damascene processing, have not been described in detail to avoid unnecessarily obscuring embodiments of the present disclosure. Furthermore, it should be understood that the various embodiments shown in the figures are illustrative representations and are not necessarily drawn to scale. In some instances, various operations will be described as multiple discrete operations, in turn, in a manner that is most helpful in understanding the present disclosure, however, the order of description should not be construed as to imply that these operations are necessarily order dependent. In particular, these operations need not be performed in the order of presentation.

Certain terminology is also used in the following description for reference purposes only, and thus these terms are not intended to be limiting. For example, terms such as "upper," "lower," "above," "below," "bottom," "top," and the like refer to the orientation in which reference is made in the drawings. Terms such as "front," "back," "rear," and "side," describe the orientation and/or position of portions of the component within a consistent but arbitrary frame of reference as may be clearly understood by reference to the text and associated drawings describing the component in question. Such terminology may include the words specifically mentioned above, derivatives thereof, and words of similar import.

Embodiments described herein may relate to front end of line (FEOL) semiconductor processing and structures. FEOL is the first part of Integrated Circuit (IC) fabrication in which individual devices (e.g., transistors, capacitors, resistors, etc.) are patterned in a semiconductor substrate or semiconductor layer. FEOL generally covers all processes up to (but not including) the deposition of metal interconnect layers. The result is typically a wafer with isolated transistors (e.g., without any conductive lines) immediately after the final FEOL operation.

Embodiments described herein may relate to back end of line (BEOL) semiconductor processing and structures. BEOL is the second part of IC fabrication, in which individual devices (e.g., transistors, capacitors, resistors, etc.) are interconnected using wiring (e.g., one or more metallization layers) on a wafer. The BEOL includes contacts, insulating layers (dielectrics), metal levels, and bonding sites for chip-to-package connections. In the BEOL portion of the fabrication stage, contacts (pads), interconnect lines, vias and dielectric structures are formed. For modern IC processes, more than 10 metal layers may be added in the BEOL.

The embodiments described below may be applicable to FEOL processes and structures, BEOL processes and structures, or to both FEOL and BEOL processes and structures. In particular, although exemplary processing schemes may be illustrated using FEOL processing scenarios, such schemes may be equally applicable to BEOL processing. In particular, although exemplary processing schemes may be illustrated using BEOL processing scenarios, such schemes may be equally applicable to FEOL processing.

One or more embodiments of the invention relate to Magnetic Tunnel Junction (MTJ) integration on the back side of silicon. Common applications for such arrays include, but are not limited to, embedded memories, magnetic tunnel junction architectures, MRAM, non-volatile memory, spin hall effect, spin torque memory, and embedded memories using magnetic storage devices.

More specifically, one or more embodiments of MTJ (magnetic tunnel junction) MRAM (magnetic random access memory) bit cells with Metallization (MOBS) on both sides are described. In one embodiment, the bitcell is fabricated on two sides of a substrate having a front side and a back side, with the first conductive line on the back side and the second conductive line on the front side. The transistor is between the second conductor and the substrate on the front side. A Magnetic Tunnel Junction (MTJ) is between the first conductive line and the substrate on the backside, wherein one end of the MTJ is coupled to the transistor through the substrate and an opposite end of the MTJ is connected to the first conductive line, and wherein the transistor is further connected to the second conductive line on the front side. In one embodiment, the bit line is a 1T (one transistor) -1MTJ bit cell with Metal (MOBS) on both sides. Embodiments also describe a layout of a 1T-1MTJ bit cell with a MOBS.

To provide context, FIG. 1 shows a two terminal 1T-1MTJ (magnetic tunnel junction) bit cell 100 for STT-MRAM. All components are shown on the same side of the substrate. The read and write current paths of bitcell 100 are equivalent, creating many design tradeoffs. For example, it is desirable that the MTJ device have a higher resistance during a read operation than during a write operation. However, the same current path for passing the read and write currents prevents having different resistances for the read and write operations. To write a logic high to bit cell 100 and raise the bit line relative to the source line (or select line), and to write a logic low to bit cell 100, lower the bit line relative to the source line. To read from the bit cell 100, the source line is set to logic low and the MTJ resistance is sensed using a weak current (e.g., 1/8 for the write current).

The 1T-1MTJ bit cell 100 may have large write current (e.g., greater than 100 μ A) and large voltage (e.g., greater than 0.7V) requirements of the MTJ. In an MTJ-based MRAM, the 1T-1MTJ bit cell 100 may have a high write error rate or low speed switching (e.g., over 20 ns). The 1T-1MTJ bit cell 100 may also have reliability issues due to tunneling current in the magnetic tunnel junction. For example, the insulator layer in the MTJ device is a barrier (e.g., 1K Ω to 10K Ω) that blocks a large current flow, and a lower current flow may cause a higher write error.

In accordance with one or more embodiments, an improved implementation for MTJ (magnetic tunnel junction) MRAM (magnetic random access memory) is provided that involves a 1T-1MTJ MRAM bit cell with Metal (MOBS) on both sides, as shown in fig. 2A.

FIG. 2A shows an integrated circuit including a 1T-1MTJ MRAM bit cell 200 having a MOBS according to one embodiment of the disclosure. Bit cell 200 includes a substrate 202 having a metallized frontside 204 and a metallized backside 206, with a first conductive line 212 (e.g., a bit line) on the backside and a second conductive line 208 (e.g., a source line) on the frontside. The transistor 210 is located between the second conductor 208 and the substrate 202 on the front side 204. The MTJ device 214 is between the first conductive line 212 and the substrate 202 on the back side 206, where one end of the MTJ device 214 is coupled to the transistor 210 through the substrate 202 and an opposite end of the MTJ device 214 is connected to the first conductive line 212 in the back side 206. On the positive side, the transistor 210 is further connected to the second conductor 208.

The 1T-1MTJ bit cell 200 with the MOBS provides a highly compact RAM via the Giant Spin Hall Effect (GSHE) that yields high spin injection efficiency. Some non-limiting technical effects of the embodiments are: a low programming voltage (or higher current for equivalent voltages) is achieved by GSHE; lower write error rates are achieved to achieve faster MRAM (e.g., less than 10 ns); decoupling the write path and the read path to achieve faster read latency; and a low resistance write operation is achieved that allows for the injection of a higher current to obtain ultra fast switching behavior of the MTJ.

FIG. 2B illustrates an exemplary material stack 220 for a 1T-1MTJ bit cell 200 based on GSHE Spin Orbit Torque (SOT) switching, according to one embodiment of the disclosure. An exemplary MTJ stack includes a free magnetic layer (FM1), a tunneling barrier, a fixed magnetic layer (FM2), a coupling layer, a Synthetic Antiferromagnet (SAF)/pinned layer, and a top electrode including three cap metal layers.

The MTJ essentially acts as a resistor, wherein the resistance of an electrical path through the MTJ can exist in two resistance states, either "high" or "low," depending on the direction or orientation of magnetization in the free magnetic layer as well as the fixed magnetic layer. In the case where the directions of magnetization in the free magnetic layer and the fixed magnetic layer closest thereto are substantially opposite or antiparallel to each other, a high resistance state exists. In the case where the directions of magnetization in the coupled free magnetic layer and the fixed magnetic layer closest thereto are substantially aligned or parallel to each other, a low resistance state exists. It should be understood that the terms "low" and "high" are relative to each other with respect to the resistance state of the MTJ. In other words, the high resistance state is only detectable more than the low resistance state and vice versa. Thus, with the detectable difference in resistance, the low resistance state and the high resistance state can represent different bits of information (i.e., "0" or "1").

In certain aspects, and in at least some embodiments of the invention, certain terms retain certain definable meanings. For example, a "free" layer magnetic layer is a magnetic layer that stores a calculable variable. A "fixed" magnetic layer is a magnetic layer having a fixed magnetization (magnetically harder than the free magnetic layer). The free layer and the fixed layer may be ferromagnetic layers. In one embodiment, the free layer may be complex and made of two separate magnetic layers with a coupling layer located therebetween. In one embodiment, the fixed layer is complex and made of two magnets with a coupling layer located between them. In yet another embodiment, both the free layer and the fixed layer may be complex. The tunneling barrier material is a material located between the free magnetic layer and the fixed magnetic layer. The SAF/pinning layer allows for cancellation of the dipole field around the free magnetic layer. The coupling layer assists the SAF/pinning layer in pinning the fixed layer and centers the hysteresis loop by overcoming the dipole field between the fixed magnetic layer and the free magnetic layer. In one embodiment, the coupling layer may comprise Ru, Ir, W, or Ta.

A wide range of material combinations may be used for the material stack of the 1T-1MTJ bit cell 200. For example, in one embodiment, the free magnetic layer and the fixed magnetic layer may comprise Co_xFe_yB_z(cobalt, iron, boron) wherein x, y and z are integers. The tunneling barrier layer may include an oxide layer, for example, magnesium oxide (MgO). The free magnetic layer is in direct contact with the SOT electrode 222, which SOT electrode 222 may comprise GSHE metal or a heavy metal dopant made of a heavy metal with high spin-orbit coupling, e.g., beta-tantalum (beta-Ta), beta-tungsten (beta-W), Pt, Hf, Ir, Bi, and Cu doped with elements such as iridium, bismuth, and/or any of the periodic groups 3d, 4d, 5d, and 4f, 5f of the periodic table. In another embodiment, the SOT electrode 222 and the optional SAF/pinning layer may comprise a Co/antiferromagnet, Fe/antiferromagnet, Ni/antiferromagnet, MnGa/antiferromagnet, MnGeGa/antiferromagnet, Bct-Ru/antiferromagnet, and alloys thereof. In yet another embodiment, the SOT electrode 222 and the optional SAF/pinning layer may include: including Ni1-xMxGa2S4 (where M ═ Mn, Fe, Co, and Zn) and transition metal dichalcogenides/topological insulators (e.g., BiSe2, WTe2, WSe2, MoSe2, etc.)Quasi-two-dimensional triangular antiferromagnet, IrMn, PtMn, NiMn or other triangular, Kagomi, chiral or hexagonal antiferromagnet and single crystal forms thereof, or amorphous alloys of various components thereof. In one embodiment, the SOT electrodes 222 are converted to a normal, highly conductive metal (e.g., Cu) to minimize the SOT electrode resistance. In alternative embodiments, other materials may be used to form the 1T-1MTJ bit cell 200.

Fig. 2C is a top view 230 of the device of fig. 2B. In fig. 2C, the magnet is oriented along the width of the SOT electrode 222 for proper spin injection. The magnetic cell is written by applying a charge current through the SOT electrode 222. The direction of magnetic writing is determined by the direction of the applied charge current. A positive current (e.g., in + y) generates a spin injection current with a transport direction (in + z) and spins pointing in the (+ x) direction. The SOT can have an effect on both the perpendicular magnetic free layer and the in-plane magnetic free layer, and the present disclosure applies to both. Since the so-called spin hall effect may be responsible for current induced magnetization switching in MTJ devices, SOT-MRAM may also be referred to as giant spin hall effect (GSPHE) MRAM.

Fig. 2D is a sectional view of the SOT electrode 222 showing the directions of spin current and charge current determined by SOT in metal. The injected spin current then generates a spin torque to align the magnet in the + x or-x direction. The current for the charge in the SOT electrode 222 is given by equation (1)

Transverse spin current of (

Having a spin direction

)：

Wherein the content of the first and second substances,

is the spin Hall injection efficiency as the ratio of the amplitude of the transverse spin current to the lateral charge current, w is the width of the magnet, t is the thickness λ of the GSHE metal electrode_sfIs the spin flip length in the GSHE metal, and θ GSHE is the spin hall angle of the GSHE metal relative to the FM1 interface. By passing

Giving an injected spin angular momentum that causes spin torque.

FIG. 3 illustrates in more detail a memory device including a 1T-1MTJ MRAM bit cell 300 with a MOBS, where like components to those of FIG. 2A have like reference numerals, but are not limited thereto, according to one embodiment of the disclosure. MRAM bit cell 300 is fabricated on both the front side 204 and the back side 206 of substrate 202 and is configured as a three-terminal device as compared to two-terminal bit cell 100. Backside 206 includes bit line 312 and MTJ device 214, where read bit line 312a and write bit line 312b of bit line 312 are decoupled from each other, forming a first terminal and a second terminal. The positive side 204 includes a source line 308 and a transistor 210, the source line 308 forming a third terminal, the transistor 210 being controllable by a Word Line (WL)224 and coupled to the source line.

In accordance with the disclosed embodiment, the MTJ device 214 on the back side 206 includes an SOT electrode 222 comprising SHE material and a free magnetic layer (e.g., CoFeB) in direct contact with the SOT electrode 222, wherein the SOT electrode 222 defines one end of the MTJ device 214 and is directly coupled to the write bit line 312 b; and the top electrode defines an opposite end of the MTJ device 214 and is coupled to the read bit line 312 a. In one embodiment, the SOT electrode 222 is shared exclusively by the MTJ device, i.e., it is not shared with other MTJ devices.

In one embodiment, one of the drain/source terminals of the transistor 210 on the front side 204 is coupled to the SOT electrode 222 on the back side 206 through a via 226 in the substrate 202, and the other of the drain/source terminals is coupled to a source line 308 on the front side 204. In one embodiment, word line 224 is coupled to the gate terminal of transistor 210. In one embodiment, the transistor 210 is an n-type transistor (e.g., NMOS) or a p-type transistor (e.g., PMOS). In one embodiment, transistor 210 may be placed in saturation mode to overcome existing limitations in highly scaled MRAM arrays.

In one embodiment, to write data to the bit cell 200, a spin current is injected into the free magnetic layer of the MTJ device in direct contact with the SOT electrode 222 formed of the SHE material. In one embodiment, to read data from the bit cell 200, a sense amplifier (not shown) senses the read bit line 312a and the write bit line 312 b.

Bitcell 200 has several advantages over bitcell 100. For example, the write and read operations of the bit cell 200 are decoupled from each other to allow for highly optimized write operations, e.g., shorter than 10ns and with very low BER (bit error rate). For example, other advantages include: the read path resistance can now be optimized for the read sense amplifier requirements; feasibility of achieving spin injection efficiency of about 100% or higher due to spin hall enhancement; the same density as that of the existing 1T-1MTJ design.

FIG. 4 is a top view of a layout of a cross-sectional view of a 1T-1MTJ MRAM bit cell with a MOBS according to one embodiment of the present disclosure. The bitcell layout 400 shows a transistor region 402 that includes a source 404, a gate 406, and a drain 408. The source line 410 extends above the transistor region 402 and the MTJ 412 is located below a portion of the drain 408. As shown, the 1T-1MTJ bit cell layout 400 has a 1.5x transistor pitch and a 1.0x metal 0 (metal 0) pitch. The 1.5P, 1.0M0 layout is 33% tighter than the 1.5P, 1.5M0 layout of the conventional 1T-1R layout. Another advantage of the bit cell layout 400 is that the layout allows for high temperature processing of the MTJ stack because the MTJ stack is on a different side of the substrate than the logic cells.

FIG. 5 is a cross-sectional view of a 1T-1MTJ MRAM bit cell having a MOBS along line cross-sectional line AA of FIG. 4, wherein like components of FIG. 4 have like reference numerals. The cross-sectional view of the bit cell layout 400 shows the transistor region 402 fabricated on the front side 502 of the substrate 500 and the MTJ 412 fabricated on the back side 504 of the substrate 500. In one embodiment, the source 404 and drain 408 of the transistor are coupled to the metal layer TCN, and then to the M0C line and M0B line, respectively, where M0C and M0B are sections of metal in the M0 layer. In one embodiment, M0C is a continuous line for a row of bit cells in the array. In one embodiment, the source line is coupled to M0C. The drain 408 is coupled to the MTJ 412 on the back side 504 using a through via 506 in the substrate 500. The through via 506 is connected to a via pedestal 508 that is in contact with one end of the MTJ 412. The opposite end of the MTJ 412 is coupled to a bit line 510.

FIG. 6 is a graph 600 of write energy-delay conditions for a 1T-1MTJ MRAM bit cell compared to a conventional MTJ according to one embodiment. The x-axis is energy (fJ/write) and the y-axis is delay in nanoseconds. Graph 700 shows five waveforms. Graph 700 compares energy-delay trajectories of GSHE and MTJ (GSHE-MTJ) devices for in-plane magnet switching as applied write voltage varies. The energy-delay trajectory (for in-plane switching) can be written as:

wherein R is_writeIs the write resistance of the device (RGSHE or RMTJ-P, RMTJ-AP), "P" is the spin current polarization (PGSHE or PMTJ), μ₀Is the vacuum permeability and "e" is the electron charge. The energy at a given delay is proportional to the square of the Gilbert damping. For various GSHE metal electrodes, τ is the change in spin polarization₀＝M_sVe/I_cPμ_BChanges also occur. The combined effect of spin hall polarization, damping and resistivity of the spin hall electrode is plotted in graph 600.

All cases considered in graph 600 assume a 30 x 60nm magnet with a 40kT thermal energy barrier and a GSHE electrode thickness of 3.5 nm. Assuming a voltage sweep according to the voltage limits of the scaled CMOS is 0V-0.7V, an energy-delay trajectory of the device is obtained. The energy-delay trajectory of a GSHE-MTJ device exhibits two regions of operation in general. In region 1, the energy delay product is approximately constant, expressed as:

(τ_d<M_sVe/I_cPμ_B) (4)

in region 2, the energy is proportional to the delay, which is expressed as:

τ_d>M_sVe/I_cPμ_B (5)

the two regions are separated by an energy minimum at:

τ_opt＝M_sVe/I_cPμ_B (6)

wherein a minimum switching energy is obtained for a spin torque device.

The energy-delay trajectory (graphs 604 and 605) of the STT-MTJ device is limited to have a minimum delay of 1ns for the in-plane device at 0.7V maximum applied voltage, with the switching energies of P-AP and AP-P being in the range of 1 pJ/write. In contrast, the energy-delay trajectory of GSHE-MTJ (in-plane anisotropy) devices 701, 702, and 703 can achieve switching times as low as 20ps (with 0.7V β -W, 20 fJ/bit) or as small as 2fJ (with 0.1V β -W, 1.5ns switching time). Graph 700 shows that 1T-1SHE MTJ devices exhibit lower write operation delay with the same energy.

FIG. 7 is a graph 700 of reliable write times for a 1T-1MTJ MRAM bit cell with a MOBS and a conventional MTJ, according to one embodiment. Graph 700 shows the write time of a simulated 1T-1SHE MTJ device using a bitcell circuit kinematically coupled with a Landau-Lifshitz-Gilbert nanomagnet. Spin hall MTJs exhibit significant write time improvements compared to perpendicular MTJs and in-plane MTJs.

FIG. 8 is a flow chart representing various operations in a method of fabricating a 1T-1MTJ memory device having a MOBS according to embodiments disclosed herein. As previously described, the fabrication techniques of the 1T-1MTJ memory device will be implemented in the context of a MOBS scheme. In some such embodiments, the MOBS scheme may be implemented by forming a first multilayer substrate and a second multilayer substrate comprising a bulk wafer (e.g., bulk silicon) or a semiconductor-on-insulator wafer (e.g., silicon-on-insulator or SOI wafer) (block 800).

Standard front-end processing may then be performed on the first substrate to form as many semiconductor devices (e.g., transistors) as desired (block 802).

Standard back-end processing may then be performed over the transistors to form contacts and as many metal (or otherwise conductive) back-end layers as desired on the first substrate (block 804). In some embodiments, the front side via or contact may be processed very deep, for example, into at least a portion of the substrate below the device layer, as the deep processed via may be used to make through-wafer contact between the drain of the transistor and the MTJ.

Thereafter, standard front end processing may be performed on the second substrate as many semiconductor devices (e.g., MTJs) are formed as desired (block 806). In an embodiment, the SOI electrode of the MTJ is formed in the dielectric layer by a damascene or dual damascene process as is well known in the art. In an embodiment, the SOI electrode may comprise a Giant Spin Hall Effect (GSHE) metal made of beta-tantalum (beta-Ta), beta-tungsten (beta-W), Pt, copper (Cu) doped with an element such as iridium, bismuth, and any of the elements in the 3d, 4d, 5d, and 4f, 5f periodic groups of the periodic table.

In one embodiment, a MTJ material stack is formed on an SOI electrode. In one embodiment, the MTJ material stack and the material layer stack are blanket deposited. The layers of the MTJ stack may be formed by a sputter deposition technique at a deposition rate

Within the range. Such techniques include Physical Vapor Deposition (PVD), especially planar magnetron sputtering, and ion beam deposition. In an embodiment, the MTJ stack may be subjected to an annealing process performed at a temperature between 300 degrees celsius and 400 degrees celsius. In an embodiment, the layers of the material layer stack may be blanket deposited by an evaporation process, an Atomic Layer Deposition (ALD) process or a Chemical Vapor Deposition (CVD) process, respectively. In an embodiment, the chemical vapor deposition process is augmented by a plasma technique (e.g., RF glow discharge)Strong (plasma enhanced CVD) to improve film density and uniformity. In an embodiment, the uppermost layer of the material layer stack may comprise a top electrode layer which eventually acts as a hard mask.

The deposition process may be configured to control the magnetic properties of the magnetic layer. For example, the direction of the magnetic anisotropy of the ferromagnetic material may be set during layer deposition by applying a magnetic field across the substrate. The resulting uniaxial anisotropy is observed as a magnetic easy direction and a magnetic hard direction in the magnetization of the layer. Since the anisotropy axis affects the switching behavior of the material, the deposition system must be able to project a uniform magnetic field, typically in the range of 20-100Oe, across the substrate during deposition. The deposition process is capable of controlling other magnetic properties, such as coercivity and magnetostriction, by selecting the magnetic alloy and deposition conditions. Since the switching field of the patterned bit depends directly on the thickness of the free layer magnet, thickness uniformity and repeatability must meet stringent requirements.

Standard back end processing may then be performed over the MTJ to form contacts and as many metal (or otherwise conductive) back end layers as desired on the second substrate (block 808). For example, a bit line may be patterned on the uppermost surface of the top electrode of the MTJ to complete the formation of the memory cell. In an embodiment, the bit line may include a conductive material such as W, TiN, TaN, or Ru. In an embodiment, the bit lines are formed by using a dual damascene process (not shown) and include a barrier layer such as Ru, Ta, or Ti and a fill metal such as W or Cu.

Thereafter, the MTJ from the second substrate is attached to the first substrate (block 810). In one exemplary process flow, this may be accomplished as follows. The MTJ stack is formed over a transfer layer on a second substrate. Then, a temporary substrate is formed on top of the second substrate. Thereafter, the second substrate is separated from the MTJ stack at the transfer layer. Thereafter, the MTJ stack with the temporary substrate thereon is attached to a first substrate having device layers including transistors formed thereon. The transfer layer of the temporary wafer is then removed, for example by etching.

In one embodiment, the MTJ from the second substrate may be attached to the first substrate prior to back end processing (block 808). In another embodiment, the MTJ may be fabricated on a first substrate and the transistor fabricated on a second substrate. In embodiments where the substrate is a wafer, respective dies from the first wafer and the second wafer may be bonded together. The die may be bonded using any suitable wafer bonding process known to those of ordinary skill in the art.

The transistors in each 1T-1MTJ bit cell with MOBS are connected to word lines and source lines in a manner that will be understood by those skilled in the art. The 1T-1MTJ bit cell with the MOBS can also include additional read and write circuitry (not shown), a sense amplifier (not shown), a bit line reference (not shown), etc., for operation of the 1T-1MTJ bit cell with the MOBS, as will be understood by those skilled in the art. It should be appreciated that a plurality of 1T-1MTJ bit cells with MOBS are operatively connected to one another to form a memory array (not shown), wherein the memory array may be incorporated into a non-volatile memory device.

Suitable semiconductor substrates include, but are not limited to, single crystal silicon, polycrystalline silicon, and silicon-on-insulator (SOI) and similar substrates formed of other semiconductor materials. Depending on the manufacturing stage, semiconductor substrates often include transistors, integrated circuits, and the like. The substrate may also include semiconductor materials, metals, dielectrics, dopants, and other materials commonly found in semiconductor substrates. Furthermore, although not shown, the structures described herein may be fabricated on an underlying lower-level back-end-of-line (BEOL) interconnect layer. For example, in one embodiment, the embedded non-volatile memory structure is formed on a material composed of a dielectric material, such as, but not limited to, silicon dioxide, silicon oxynitride, silicon nitride, or carbon-doped silicon nitride.

Referring to fig. 9A and 9B, a wafer 900 may be composed of a semiconductor material and may include one or more dies 902 having Integrated Circuit (IC) structures formed on a surface of the wafer 900. Each of the dies 902 may be a repeating unit of a semiconductor product that includes any suitable IC, e.g., an IC that includes one or more embedded non-volatile memory structures having 1T-1MTJ memory devices with MOBSs, e.g., as described above. After fabrication of the semiconductor product is complete, the wafer 900 may undergo a singulation process in which each of the dies 902 is separated from one another to provide discrete "chips" of the semiconductor product. In particular, the structure including the embedded non-volatile storage structure having the 1T-1MTJ storage device with MOBS disclosed herein may be in the form of a wafer 900 (e.g., not singulated) or in the form of a die 902 (e.g., singulated). Die 902 may include one or more embedded non-volatile memory structures having 1T-1MTJ memory devices with MOBS and/or supporting circuitry to route electrical signals, as well as any other IC components. In some embodiments, wafer 900 or die 902 may include additional memory devices (e.g., Static Random Access Memory (SRAM) devices), logic devices (e.g., and, or, nand, or nor gates), or any other suitable circuit elements. Multiple ones of these devices may be combined on a single die 902. For example, a memory array formed by a plurality of memory devices may be formed on the same die 902 as a processing device or other logic configured to store information in the memory devices or execute instructions stored in the memory array.

The embodiments disclosed herein may be used to fabricate a wide variety of different types of integrated circuits and/or microelectronic devices. Examples of such integrated circuits include, but are not limited to, processors, chipset components, graphics processors, digital signal processors, microcontrollers, and the like. In other embodiments, semiconductor memories may be fabricated. Furthermore, integrated circuits or other microelectronic devices may be used in a wide variety of electronic devices known in the art. For example, in computer systems (e.g., desktop, laptop, server), cellular phones, personal electronics. The integrated circuit may be coupled to a bus and other components in the system. For example, the processor may be coupled to the memory, chipset, etc. by one or more buses. It is possible to manufacture each of the processors, memories, and chipsets using the approaches disclosed herein.

Fig. 10 is a cross-sectional side view of an Integrated Circuit (IC) device assembly that may include one or more embedded non-volatile memory structures having a 1T-1MTJ memory device with a MOBS in accordance with one or more of the embodiments disclosed herein.

Referring to fig. 10, an IC device assembly 1000 includes components having one or more integrated circuit structures described herein. The IC device assembly 1000 includes several components disposed on a circuit board 1002 (which may be, for example, a motherboard). The IC device assembly 1000 includes components disposed on a first side 1040 of the circuit board 1002 and an opposing second side 1042 of the circuit board 1002. In general, features may be disposed on one or both of faces 1040 and 1042. In particular, any suitable ones of the components of the IC device assembly 1000 may include an embedded non-volatile memory structure having a 1T-1MTJ memory device with a MOBS, e.g., as disclosed herein.

In some embodiments, circuit board 1002 may be a Printed Circuit Board (PCB) that includes multiple metal layers separated from each other by layers of dielectric material and interconnected by conductive vias. Any one or more of the metal layers may be formed in a desired circuit pattern to route electrical signals between components coupled to the circuit board 1002 (optionally in conjunction with other metal layers). In other embodiments, the circuit board 1002 may be a non-PCB substrate.

The IC device assembly 1000 shown in fig. 10 includes an on-interposer package structure 1036 coupled to the first side 1040 of the circuit board 1002 by coupling members 1016. Coupling components 1016 may electrically and mechanically couple on-interposer package 1036 to circuit board 1002 and may comprise solder balls (as shown in fig. 8), male and female portions of a socket, adhesive, underfill material, and/or any other suitable electrical and/or mechanical coupling structure.

The package-on-interposer structure 1036 may include an IC package 1020 coupled to the interposer 1004 by a coupling member 1018. The coupling component 1018 may take any suitable form for the application, such as the form discussed above with reference to the coupling component 1016. Although fig. 10 shows a single IC package 1020, multiple IC packages may be coupled to interposer 1004. It should be appreciated that additional interpolators may be coupled to interpolator 1004. Interposer 1004 may provide an intervening substrate for bridging circuit board 1002 and IC package 1020. IC package 1020 may be or may include, for example, a die (die 902 of fig. 9B) or any other suitable component. In general, interposer 1004 may expand a connection to a wider pitch or reroute a connection to a different connection. For example, the interposer 1004 may couple the IC package 1020 (e.g., die) to a Ball Grid Array (BGA) of the coupling component 1016 for coupling to the circuit board 1002. In the embodiment shown in fig. 10, IC package 1020 and circuit board 1002 are attached to opposite sides of interposer 1004. In other embodiments, IC package 1020 and circuit board 1002 may be attached to the same side of interposer 1004. In some embodiments, three or more components may be interconnected by interposer 1004.

Interposer 1004 may be formed of epoxy, fiberglass-reinforced epoxy, ceramic material, or polymer material such as polyimide. In some implementations, interposer 1004 can be formed of alternating rigid or flexible materials, which can include the same materials as described above for use in semiconductor substrates, e.g., silicon, germanium, and other group III-V and group IV materials. Interposer 1004 may include metal interconnects 1010 and vias 1008, vias 1008 including, but not limited to, Through Silicon Vias (TSVs) 1006. Interposer 1004 may also include embedded devices, including both passive and active devices. Such devices may include, but are not limited to, capacitors, decoupling capacitors, resistors, inductors, fuses, diodes, transformers, sensors, electrostatic discharge (ESD) devices, and memories. More complex devices such as Radio Frequency (RF) devices, power amplifiers, power management devices, antennas, arrays, sensors, and micro-mechanical systems (MEMS) devices may also be formed on interposer 1004. The on-interposer package structure 1036 may be in the form of any of the on-interposer package structures known in the art.

The IC device assembly 1000 may include an IC package 1024 coupled to a first side 1040 of the circuit board 1002 by a coupling component 1022. The coupling component 1022 may take the form of any of the embodiments discussed above with reference to the coupling component 1016, and the IC package 1024 may take the form of any of the embodiments discussed above with reference to the IC package 1020.

The IC device assembly 1000 shown in fig. 10 includes a package-on-package structure 1034 coupled to a second side 1042 of the circuit board 1002 by a coupling member 1028. The package-on-package structure 1034 may include an IC package 1026 and an IC package 1032 coupled together by a coupling component 1030 such that the IC package 1026 is disposed between the circuit board 1002 and the IC package 1032. The coupling components 1028 and 1030 may take the form of any of the embodiments of the coupling component 1016 discussed above, and the IC packages 1026 and 1032 may take the form of any of the embodiments of the IC package 1020 discussed above. Package-on-package structure 1034 may be configured in accordance with any package-on-package structure known in the art.

FIG. 11 illustrates a computing device 1100 according to one implementation of the disclosure. The computing device 1100 houses a board 1102. Board 1102 may include several components including, but not limited to, a processor 1104 and at least one communication chip 1106. The processor 1104 is physically and electrically coupled to the board 1102. In some embodiments, at least one communication chip 1106 is also physically and electrically coupled to board 1102. In other implementations, the communication chip 1106 is part of the processor 1104.

Depending on its application, the computing device 1100 may include other components that may or may not be physically and electrically coupled to the board 1102. These other components include, but are not limited to, volatile memory (e.g., DRAM), non-volatile memory (e.g., ROM), flash memory, a graphics processor, a digital signal processor, a crypto processor, a chipset, an antenna, a display, a touchscreen controller, a battery, an audio codec, a video codec, a power amplifier, a Global Positioning System (GPS) device, a compass, an accelerometer, a gyroscope, a speaker, a camera, and a mass storage device (e.g., hard disk drive, Compact Disc (CD), Digital Versatile Disc (DVD), etc.).

The communication chip 1106 enables wireless communication of data to and from the computing device 1100. The term "wireless" and its derivatives may be used to describe circuits, devices, systems, methods, techniques, communications channels, etc., that may communicate data through the use of modulated electromagnetic radiation through a non-solid medium. The term does not imply that the associated devices do not contain any wires, but in some embodiments they may not. The communication chip 1106 may implement any of a number of wireless standards or protocols, including but not limited to Wi-Fi (IEEE 802.11 family), WiMAX (IEEE 802.16 family), IEEE802.20, Long Term Evolution (LTE), Ev-DO, HSPA +, HSDPA +, HSUPA +, EDGE, GSM, GPRS, CDMA, TDMA, DECT, Bluetooth, derivatives thereof, and any other wireless protocol known as 3G, 4G, 5G, and higher generation. The computing device 1100 may include a plurality of communication chips 1106. For example, a first communication chip 1106 may be dedicated to shorter range wireless communications, such as Wi-Fi and Bluetooth, and a second communication chip 1106 may be dedicated to longer range wireless communications, such as GPS, EDGE, GPRS, CDMA, WiMAX, LTE, Ev-DO, and others.

The processor 1104 of the computing device 1100 includes an integrated circuit die packaged within the processor 1104. In some implementations of the disclosure, an integrated circuit die of a processor includes one or more embedded non-volatile storage structures having a 1T-1MTJ memory device with a MOBS in accordance with implementations of embodiments of the disclosure. The term "processor" may refer to any device or portion of a device that processes electronic data from registers and/or memory to transform that electronic data into other electronic data that may be stored in registers and/or memory.

The communication chip 1106 also includes an integrated circuit die packaged within the communication chip 1106. According to another implementation of an embodiment of the present disclosure, an integrated circuit die of a communication chip includes one or more embedded non-volatile storage structures having 1T-1MTJ memory devices with MOBSs according to an implementation of an embodiment of the present disclosure.

In other implementations, another component housed within the computing device 1100 may include an integrated circuit die that includes one or more embedded non-volatile storage structures having 1T-1MTJ memory devices with MOBSs in accordance with implementations of embodiments of the present disclosure.

In various implementations, the computing device 1100 may be a laptop, a netbook, a notebook, an ultrabook, a smartphone, a tablet, a Personal Digital Assistant (PDA), an ultra mobile PC, a mobile phone, a desktop computer, a server, a printer, a scanner, a monitor, a set-top box, an entertainment control unit, a digital camera, a portable music player, or a digital video recorder. In other implementations, the computing device 1100 may be any other electronic device that processes data.

Thus, embodiments described herein include embedded non-volatile memory structures having 1T-1MTJ memory device elements with MOBSs

The above description of illustrated embodiments of the disclosure, including what is described in the abstract, is not intended to be exhaustive or to limit the disclosure to the precise forms disclosed. While specific embodiments of, and examples for, the disclosure are described herein for illustrative purposes, various equivalent modifications are possible within the scope of the disclosure, as those skilled in the relevant art will recognize.

These modifications can be made to the present disclosure in light of the above detailed description. The terms used in the following claims should not be construed to limit the disclosure to the specific embodiments disclosed in the specification and the claims. Rather, the scope of the disclosure is to be determined entirely by the following claims, which are to be construed in accordance with established doctrines of claim interpretation.

The following examples relate to other embodiments. Various features of the different embodiments may be combined in various ways, including some features and excluding other features, to accommodate a wide variety of different applications.

Exemplary embodiment 1: a memory device includes a substrate having a front side and a back side, wherein first conductive lines are on the back side and second conductive lines are on the front side. A transistor is between the second conductive line and a substrate on the front side. A Magnetic Tunnel Junction (MTJ) is between a first conductive line and a substrate on a backside, wherein one end of the MTJ is coupled to the transistor through the substrate and an opposite end of the MTJ is connected to the first conductive line, and wherein the transistor is further connected to a second conductive line on the front side.

Exemplary embodiment 2: the memory device of example embodiment 1, wherein an MTJ is connected to the drain of the transistor and the source of the transistor is coupled to the second conductive line.

Exemplary embodiment 3: the memory device of claim 1 or 2, wherein the MTJ is connected to the drain of the transistor using a through via extending through the substrate from the front side to the back side.

Exemplary embodiment 4: the memory device of claim 3, wherein a gate of the transistor is coupled to a word line.

Example embodiment 5 the memory device of claim 1, 2 or 3, wherein the first conductive line comprises a bit line and the second conductive line comprises a source line.

Exemplary embodiment 6: the memory device of claim 5, wherein the bit lines include a read bit line and a write bit line.

Exemplary embodiment 7: the memory device of claim 1, 2, 3, 4, 5, or 6, wherein the memory device comprises a 1T-1MTJ Magnetic Random Access Memory (MRAM).

Exemplary embodiment 8: the memory device of claim 1, 2, 3, 4, 5, 6, or 7, wherein the MTJ device comprises an SOT electrode comprising GSHE material.

Exemplary embodiment 9: the memory device of claim 8, wherein the GSHE material includes at least one of β -tantalum (β -Ta), β -tungsten (β -W), Pt, Hf, Ir, Bi, and doped Cu.

Exemplary embodiment 10: the memory device of claim 1, 2, 3, 4, 5, 6, 7, 8, or 9, wherein the MTJ device comprises a free magnetic layer coupled with the SOT electrode, wherein the SOT electrode is coupled to a write bit line; and an opposite end of the MTJ device is coupled to a read bit line.

Exemplary embodiment 11: the memory device of claim 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10, wherein the material stack comprising the MTJ device further comprises: a tunneling barrier, a fixed magnetic layer, a coupling layer, a Synthetic Antiferromagnet (SAF)/pinning layer, and a top electrode.

Exemplary embodiment 12: a memory device includes a substrate. The backside of the substrate includes read and write bit lines and a Magnetic Tunnel Junction (MTJ) device. The front side of the substrate includes a source line, and a transistor controllable by a word line and coupled to the source line.

Exemplary embodiment 13: the memory device of claim 12, wherein the MTJ device on the back side comprises a write electrode comprising a spin hall effect material and a free magnetic layer in direct contact with an SOT electrode, wherein the SOT electrode defines one end of the MTJ device and is coupled to the write bit line; and a top electrode defines an opposite end of the MTJ device and is coupled to the read bit line.

Exemplary embodiment 14: the memory device of claim 12 or 13, wherein one of the drain/source terminals of the transistor on the front side is coupled to the SOT electrode on the back side through a via in the substrate.

Exemplary embodiment 15: the memory device of claim 14, wherein the through via is connected to a via pedestal in contact with one end of the MTJ device.

Exemplary embodiment 16: the memory device of claim 12, 13, 14 or 15, wherein the memory device comprises a three terminal device, wherein read and write bit lines on the backside form a first terminal and a second terminal, and the source line forms a third terminal.

Exemplary embodiment 17: the memory device of claim 12, 13, 14, 15 or 16, wherein one of the drain/source terminals of the transistor on the positive side is coupled through the substrate to the SOT electrode on the back side and the other of the source/drain terminals is coupled to the source line on the positive side.

Exemplary embodiment 18: the memory device of claim 12, 13, 14, 15, 16, or 17, wherein the word line is coupled to a gate terminal of a transistor.

Exemplary embodiment 19: the memory device of claim 12, 13, 14, 15, 16, 17, or 18, wherein the memory device comprises a 1T-1MTJ Magnetic Random Access Memory (MRAM).

Exemplary embodiment 20: the memory device of claim 12, 13, 14, 15, 16, 17, 18, or 19, wherein the MTJ device comprises an SOT electrode comprising GSHE material.

Exemplary embodiment 21: the memory device of claim 20, wherein the GSHE material includes at least one of β -tantalum (β -Ta), β -tungsten (β -W), Pt, Hf, Ir, Bi, and doped Cu.

Exemplary embodiment 22: the memory device of claim 19 or 20, wherein the MTJ device comprises a free magnetic layer coupled with the SOT electrode, wherein the SOT electrode is coupled to a write bit line; and an opposite end of the MTJ device is coupled to a read bit line.

Exemplary embodiment 23: the memory device of claim 22, wherein the material stack comprising the MTJ device further comprises: a tunneling barrier, a fixed magnetic layer, a coupling layer, a Synthetic Antiferromagnet (SAF)/pinning layer, and a top electrode.

Exemplary embodiment 24: a method of manufacturing an integrated circuit device includes forming a first substrate and a second substrate. Front end processing is performed on the first substrate to form a transistor. Back-end processing is performed over the transistors to form first contacts and conductive back-end layers on the first substrate. Performing front-end processing on the second substrate to form a Magnetic Tunnel Junction (MTJ), performing back-end processing over the MTJ to form a second contact and a conductive back-end layer on the second substrate. Thereafter, the MTJ from the second substrate is attached to the first substrate.

Exemplary embodiment 25: the method of claim 24, further comprising: coupling one of the drain/source terminals of the transistor on the front side of the first substrate to the MTJ on the back side of the first substrate using a through via in the first substrate.