THE SAME

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefits of U.S. Patent Application No. 17/658,641, entitled “Resistive Random- Access Memory Devices with Engineered Electronic Defects and Methods for Making the Same,” filed April 8, 2022, which claims the benefits of U.S. Patent Application No. 17/454 ,914, entitled “Resistive Random-Access Memory Devices with Multi- Component Electrodes and Discontinuous Interface Layers,” filed November 15, 2021,

U.S. Patent Application No. 17/319,057, entitled “Resistive Random-Access Memory Devices with Multi-Component Electrodes,” filed May 12, 2021, U.S. Patent Application No.

17/319,068, entitled “Resistive Random- Access Memory Devices with Multi-Component Electrodes,” filed May 12, 2021, and PCT Application No. PCT/US21/40389, entitled “Resistive Random- Access Memory Devices with Multi-Component Electrodes,” filed July 03, 2021, each of which is incorporated herein in its entirety.

TECHNICAL FIELD

[0002] The implementations of the disclosure relate generally to resistive random-access memory (RRAM) devices and, more specifically, to RRAM devices with engineered electronic defects and methods for fabricating the same.

BACKGROUND

[0003] A resistive random-access memory (RRAM) device is a two-terminal passive device with tunable and non-volatile resistance. The resistance of the RRAM device may be electrically switched between a high-resistance state (HRS) and a low-resistance state (LRS) by applying suitable programming signals to the RRAM device. RRAM devices may be used to form crossbar arrays that may be used to implement in-memory computing applications, non volatile solid-state memory, image processing applications, neural networks, etc.

SUMMARY

[0004] The following is a simplified summary of the disclosure in order to provide a basic understanding of some aspects of the disclosure. This summary is not an extensive overview of the disclosure. It is intended to neither identify key or critical elements of the

1 disclosure, nor delineate any scope of the particular implementations of the disclosure or any scope of the claims. Its sole purpose is to present some concepts of the disclosure in a simplified form as a prelude to the more detailed description that is presented later.

[0005] One or more aspects of the present disclosure provide a method for fabricating a resistive random-access memory (RRAM) device. The method includes fabricating, on a first electrode of the RRAM device, a first interface layer comprising a first discontinuous film of a first material; fabricating, on the first interface layer, a switching oxide layer comprising at least one transition metal oxide; fabricating, on the switching oxide layer, a second interface layer comprising a second discontinuous film of a second material; and fabricating, on the second interface layer, a defect engineering layer for generating electronic defects in the switching oxide layer. The first material and the second material are more chemically stable than the at least one transition metal oxide.

[0006] In some embodiments, the at least one transition metal oxide includes at least one of HfOx or TaOy, wherein x<2.0, and wherein y<2.5.

[0007] In some embodiments, the first material includes at least one of AI2O3, MgO,

Y2O3, or La203.

[0008] In some embodiments, fabricating on the first electrode of the RRAM device, the first interface layer including the first discontinuous film of the first material includes depositing the first material on the first electrode to form the first discontinuous film.

[0009] In some embodiments, a thickness of the first interface layer is between 0.2nm and lnm.

[0010] In some embodiments, the second material includes at least one of AI2O3, MgO,

Y2O3, or La203.

[0011] In some embodiments, fabricating the defect engineering layer on the second interface layer includes fabricating a first layer of a first metallic material; and fabricating, on the first layer of the first metallic material, a second layer of a second metallic material.

[0012] In some embodiments, the first material is more chemically stable than the at least one transition metal oxide. The first material is more chemically stable than an oxide of the first metallic material.

[0013] In some embodiments, the first metallic material includes at least one of Ti, Hf, or

Zr.

[0014] In some embodiments, the second metallic material includes tantalum.

2 [0015] In some embodiments, fabricating the first layer of the first metallic material includes depositing a layer of Ti on the second interface layer.

[0016] In some embodiments, the second layer of the second metallic material includes one or more alloys containing tantalum.

[0017] In some embodiments, the defect engineering layer includes one or more alloys containing tantalum.

[0018] In some embodiments, the one or more alloys containing tantalum further includes at least one of hafnium, molybdenum, tungsten, niobium, or zirconium.

[0019] In some embodiments, the one or more alloys containing tantalum include at least one of a binary alloy including tantalum, a ternary alloy including tantalum, a quaternary alloy including tantalum, a quinary alloy including tantalum, a senary alloy including tantalum, or a high order alloy including tantalum.

[0020] In some embodiments, a thickness of the first layer including the first metallic material is between 0.2nm and 5nm.

[0021] In some embodiments, the defect engineering layer contacts at least a portion of the switching oxide layer.

BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The disclosure will be understood more fully from the detailed description given below and from the accompanying drawings of various embodiments of the disclosure. The drawings, however, should not be taken to limit the disclosure to the specific embodiments, but are for explanation and understanding.

[0023] FIG. 1 is a schematic diagram illustrating an example of a crossbar circuit in accordance with some implementations of the disclosure.

[0024] FIG. 2 is a schematic diagram illustrating an example of a cross-point device in accordance with some implementations of the disclosure.

[0025] FIGS. 3A-3C, 4A-4F, 5A-5D, and 6A-6D illustrate cross-sectional views of example RRAM devices in accordance with some embodiments of the present disclosure.

[0026] FIG. 7 is a schematic diagram illustrating cross-sectional views of a top electrode of an RRAM device in accordance with some embodiments of the present disclosure.

[0027] FIG. 8 is an example Ellingham diagram showing chemical stability of metal oxides in accordance with some embodiments of the present disclosure.

3 [0028] FIG. 9 depicts a tantalum-titanium (Ta-Ti) binary phase diagram in accordance with some embodiments of the present disclosure.

[0029] FIG. 10 is a flow diagram illustrating a method for fabricating an RRAM device in accordance with some embodiments of the present disclosure. [0030] FIG. 11 is a flow diagram illustrating a method for fabricating an RRAM device in accordance with some embodiments of the present disclosure.

[0031] FIG. 12 is a flow diagram illustrating a method for fabricating an RRAM device in accordance with some embodiments of the present disclosure.

[0032] FIG. 13 is a flow diagram illustrating a method for fabricating a defect engineering layer for generating electronic defects in a switching oxide layer of an RRAM device in accordance with some embodiments of the present disclosure.

[0033] FIG. 14A depicts a tantalum-hafnium (Ta-Hf) binary phase diagram in accordance with some embodiments of the present disclosure.

[0034] FIG. 14B depicts a tantalum-tungsten (Ta-W) binary phase diagram in accordance with some embodiments of the present disclosure.

[0035] FIG. 14C depicts a tantalum-molybdenum (Ta-Mo) binary phase diagram in accordance with some embodiments of the present disclosure.

[0036] FIG. 14D depicts a tantalum-niobium (Ta-Nb) binary phase diagram in accordance with some embodiments of the present disclosure. [0037] FIG. 14E depicts a tantalum-zirconium (Ta-Zr) binary phase diagram in accordance with some embodiments of the present disclosure.

[0038] FIGS. 15A, 15B, 15C, and 15D depict current-voltage characteristics of example

RRAM devices in accordance with some embodiments of the present disclosure.

4 DETAILED DESCRIPTION

[0039] Aspects of the disclosure provide resistive random-access memory (RRAM) devices and methods for fabricating the RRAM devices. An RRAM device is a two-terminal passive device with tunable resistance. The RRAM device may include a first electrode, a second electrode, and a switching oxide layer positioned between the first electrode and the second electrode. The first electrode may include a non-reactive metal, such as platinum (Pt), palladium (Pd), etc. The second electrode may include a reactive metal, such as tantalum (Ta). The electrode including the non-reactive metal is also referred to herein as the “non-reactive electrode.” The electrode including the reactive metal is also referred to herein as the “reactive electrode.” The switching oxide layer may include a transition metal oxide, such as hafnium oxide (HfO _x ) or tantalum oxide (TaO _x ). The RRAM device may be in an initial state or virgin state and may have an initial high resistance before it is subject to a suitable electrical stimulation (e.g., a voltage or current signal applied to the RRAM device). The RRAM device may be tuned to a lower resistance state from the virgin state via a forming process or from a high-resistance state (HRS) to a lower resistance state (LRS) via a setting process. The forming process may refer to programming a device starting from the virgin state. The setting process may refer to programming a device starting from the high resistance state (HRS). After the reactive metal electrode is deposited on the switching oxide, the reactive metal can absorb oxygen from the switching oxide layer and create oxygen vacancies in the switching oxide layer, and oxygen ions can migrate in the switching oxide through a vacancy mechanism. During a forming process, a suitable programming signal (e.g., a voltage or current signal) may be applied to the RRAM device, which may cause a drift of oxygen ions to migrate from the switching oxide to the reactive electrode. As a result, a conductive channel or filament may form through the switching oxide layer (e.g., from the reactive electrode to the non-reactive electrode). The RRAM device may then be reset to a high-resistance state by applying a reset signal (e.g., a voltage signal, a current signal) to the RRAM device. The application of the reset signal to the RRAM device may cause oxygen to migrate back to the switching oxide layer and may thus interrupt the conductive filament. The RRAM device may be electrically switched between a high-resistance state and a low-resistance state by applying suitable programming signals (e.g., voltage signals, current signals, etc.) to the RRAM device. In a crossbar array circuit, the programming signals may be provided to the designated RRAM device via a selector, such as a transistor.

[0040] An RRAM device is regarded as being operated in a filamentary mode when a conductive channel or filament is formed through the switching oxide (e.g., from the reactive

5 electrode to the non-reactive electrode). The RRAM device is regarded as being operated in a non- filamentary mode when an interrupted filament is formed in the switching oxide layer of the RRAM device. There may be a gap between the interrupted filament and the bottom electrode of the RRAM device.

[0041] According to the concept of conductance quantum, a material’s electrical conductance is observed to change in discrete or quantized steps. The unit of conductance of quantum is Go=2e ² /h=7.748E-5 Siemens (or 12.9kQ), where e and h represent the electron charge and the Planck constant, respectively. The minimum quantum conductance occurs when 2 metal atoms form a point contact, which can be regarded as the minimum conductance (or maximum resistance) of a metallic filament. That is, the minimum conductance or maximum resistance in a filamentary mode is limited to Go (7.74E-5 Siemens or 12.9kG). However, the implementations of certain applications (e.g., IMC applications) may require RRAM devices with high resistance, such as a resistance higher than Go. It may be necessary to operate the RRAM devices in a non- filamentary mode to achieve such a high resistance. The electronic conduction in the devices with high resistance is thus semiconductor instead of metallic in nature. However, most RRAM devices only present certain desired features (e.g., analog resistance, multilevel resistance, linearity, etc.) when they are operated in a filamentary mode. In the filamentary mode, the electric conduction in the RRAM devices may be dominated by the conduction through the filament, while the conduction through the switching oxide is negligible. This is because the conduction band and the valance band of a metal overlap and the electrons may readily move between atoms. In a semiconductor, however, there is an energy gap (or a band gap) between the conductor band and the valance band. Electrons may need to overcome this band gap to move from the valance band to the conduction band of the semiconductor and be able to move between atoms. An electronic defect in the switching oxide layer of the RRAM device may have an energy between the valance band and the conduction band and may trap an electron that may be readily excited to the conduction band or hop from one trap site to another trap site. Therefore, for implementing device operations using RRAM devices in a non filamentary mode with high resistance, the electronic defects in the switching oxide may be important where electric conduction may be dominated by electron defects in the switching oxide where the filament may be interrupted. Accordingly, it may be desirable to engineer and control electronic defects in the switching oxides of the RRAM devices to achieve certain electronic behaviors that are critical to IMC applications, such as analog resistance, multilevel resistance, I-V (current-voltage) linearity, etc.

6 [0042] Furthermore, it might be desirable to scale down RRAM devices to a suitable size

(e.g., a critical dimension of lOOnm, lOnm, or a smaller dimension) to implement certain in memory computing (IMC) applications (e.g., an IMC application that requires high-density RRAM devices and/or low-power consumption). However, when the critical dimension of a conventional RRAM device scales down, the filament formed in the conventional RRAM device may not scale down accordingly. For example, the size of the filament formed in the scaled- down RRAM device may not be scaled down proportionally. As such, forming, setting, and/or resetting such a conventional scaled-down RRAM device may still require a relatively high current or voltage. This may also prevent the effective scaling down of the selector (e.g., a transistor) and/or the integrated circuit that provides the current or voltage to the scaled-down RRAM device. Furthermore, the scaled-down RRAM device may have a relatively smaller area of top electrode. The top electrode may not be able to absorb as much oxygen as that of a larger RRAM device. This may cause device failures and/or operation failures of the RRAM device. For example, a device failure can be caused by delamination between the reactive electrode and the switching oxide by the presence of oxygen molecules. As another example, the oxygen ions may drift from the switching oxide into the top electrode under an external voltage and may migrate back to the switching oxide once the external voltage is removed, resulting in the operation to be volatile, which is an operation failure for non-volatile memory.

[0043] Accordingly, the present disclosure provides mechanisms for engineering defects in RRAM devices that may enhance the performance of the RRAM devices and implement low- power IMC applications. In some embodiments, an RRAM device may include a bottom electrode, a first interface layer fabricated on the bottom electrode, a switching oxide layer fabricated on the first interface layer, and a top electrode. The bottom electrode may include Pt or any other suitable nonactive metal. The switching oxide layer may include a transition metal oxide, such as HfO _x , TaO _x , TiO _x , NbO _x , ZrO _x , etc. The first interface layer may include a discontinuous film of a first material that is more chemically stable than the transition metal oxide. The first material may include, for example, AI2O3, MgO, Y2O3, La2C>3, etc.

[0044] In some implementations, the RRAM device may further include a second interface layer fabricated on the switching oxide layer. In such implementations, the top electrode may be fabricated on the second interface layer. The second interface layer may include a discontinuous film of a second material that is more chemically stable than the transition metal oxide and may further restrict an electric path through the switching oxide for low current and low power operations.

7 [0045] A defect engineering layer may be fabricated on the switching oxide layer and/or the second interface layer. The defect engineering layer may include a first layer of a suitable metallic material for generating defects in the switching oxide layer. In some embodiments, the defect engineering layer may include a layer of titanium (Ti). The first layer of the metallic material may be a thin layer with a thickness between about 0.2nm and 5nm. The defect engineering layer may trap and release oxygen during device operations. The incorporation of the defect engineering layer into the RRAM device may produce a high density of electronic defects (e.g., oxygen vacancy defects) in the switching metal oxides. The electronic defects can assist charges in transporting in the switching oxide layer under an electric field at room temperature, below, or above. In some embodiments, the charges may be ionic, such as oxygen ion O ² that carries -2 charges or oxygen vacancy Vo ⁺² that carries +2 charges. In some embodiments, the charges may be electronic, such as an electron e ¹ , where e represents an electron and - 1 represents the charge carried by the electron, being trapped in the vacancy site with an energy between the valance band and conduction band. Under an electric field, the trapped electron may be excited to the conduction band (with a lower excitation energy) or hop from one trap to another trap without being excited to conduction band (also called tunneling). The incorporation of the thin defect engineering layer into the RRAM device may thus change the virgin resistance of the RRAM device, result in a less abrupt forming process, reduce the forming voltage, reduce the reset current, and reduce voltage and/or current requirements in subsequent operation processes.

[0046] The defect engineering layer may further include a second layer of Ta or any other suitable metallic material.

[0047] In some embodiments, the defect engineering layer may include one or more alloys of Ta in one implementation. The alloys of Ta may be and/or include a binary alloy containing Ta, a ternary alloy containing Ta, a quaternary alloy containing Ta, a quinary alloy containing Ta, a senary alloy containing Ta, and/or a high order alloy (e.g., an alloy containing more than six metallic elements) containing Ta. Each of the alloys may include Ta and one or more other metallic elements that have required thermodynamic and/or kinetic properties than Ta, such as tungsten (W), hafnium (Hf), molybdenum (Mo), niobium (Nb), zirconium (Zr), etc. For example, fabricating the top electrode using an alloy of Ta instead of pure Ta metal may reduce the migration of Ta into the switching oxide layer during the forming process and may thus reduce the size of the filament formed in the switching oxide layer (e.g., by reducing the lateral dimension or diameter of the filament). This may increase the filament resistance of the

8 RRAM device and may thus increase the resistance of the RRAM device in both the low- resistance state and the high-resistance state, which may thus reduce the voltage and/or current required for operations of the RRAM device, such as forming, setting, resetting, and/or tuning the RRAM device. The RRAM device incorporating the defect engineering layer may present dynamic memristive behavior in multiple dimensions suitable for implementing dynamic learning, edge processing, inference engine accelerators, and other IMC applications.

[0048] Accordingly, the present disclosure provides techniques for engineering defects in

RRAM devices to achieve high filament resistance and reduced operation voltages and currents. The RRAM devices described herein present desirable linearity, analog, retention, and endurance etc. behaviors for IMC applications in a high resistance range (e.g., from 10kQ to 10MW). The techniques may enable efficient scaling down of RRAM devices and low-power consumption IMC applications utilizing RRAM devices.

[0049] FIG. 1 is a schematic diagram illustrating an example 100 of a crossbar circuit in accordance some embodiments of the present disclosure. As shown, crossbar circuit 100 may include a plurality of interconnecting electrically conductive wires, such as one or more row wires 111a, 111b, . . ., 11 li, . . ., 11 In, and column wires 113a, 113b, . . ., 113j, . . ., 113m for an n-row by m-column crossbar array. The crossbar circuit 100 may further include cross-point devices 120a, 120b, . . ., 120z, etc. Each of the cross-point devices may connect a row wire and a column wire. For example, the cross-point device 120ij may connect the row wire 11 li and the column wire 113j. In some embodiments, crossbar circuit 100 may further include digital to analog converters (DAC, not shown), analog to digital converters (ADC, not shown), switches (not shown), and/or any other suitable circuit components for implementing a crossbar-based apparatus. The number of the column wires 113a-m and the number of the row wires 11 la-n may or may not be the same.

[0050] Row wires 111 may include a first row wire 11 la, a second row wire 111b, . . .,

11 li, . . ., and a n-th row wire 11 In. Each of row wires 11 la, . . ., 11 In may be and/or include any suitable electrically conductive material. In some embodiments, each row wire 11 la-n may be a metal wire.

[0051] Column wires 113 may include a first column wire 113a, a second column wire

113b, . . ., and an m-th column wire 113m. Each of column wires 113a-m may be and/or include any suitable electrically conductive material. In some embodiments, each column wire 113a-m may be a metal wire.

9 [0052] Each cross-point device 120 may be and/or include any suitable device with tunable resistance, such as a memristor, pulse-code modulation (PCM) devices, floating gates, spintronic devices, RRAM, static random-access memory (SRAM), etc. In some embodiments, one or more of cross-point devices 120 may include an RRAM device as described in connection with FIGS. 3A-5B.

[0053] Crossbar circuit 100 may perform parallel weighted voltage multiplication and current summation. For example, an input voltage signal may be applied to one or more rows of crossbar circuit 100 (e.g., one or more selected rows). The input signal may flow through the cross-point devices of the rows of the crossbar circuit 100. The conductance of the cross-point device may be tuned to a specific value (also referred to as a “weight”). By Ohm’s law, the input voltage multiplies the cross-point conductance generates a current form the cross-point device.

By Kirchhoff s law, the summation of the current passing the devices on each column generates the current as the output signal, which may be read from the columns (e.g., outputs of the ADCs). According to Ohm’s law and Kirchhoff s current law, the input-output relationship of the crossbar array can be represented as I=VG, wherein I represents the output signal matrix as current; V represents the input signal matrix as voltage; and G represents the conductance matrix of the cross-point devices. As such, the input signal is weighted at each of the cross-point devices by its conductance according to Ohm’ s law. The weighted current is outputted via each column wire and may be accumulated according to Kirchhoff s current law. This may enable in memory computing (IMC) via parallel multiplications and summations performed in the crossbar arrays.

[0054] FIG. 2 is a schematic diagram illustrating an example 200 of a cross-point device in accordance some embodiments of the present disclosure. As shown, cross-point device 200 may connect a bitline (BF) 211, a select line (SEE) 213, and a wordline (WE) 215. The bitline 211 and the wordline 215 may be a column wire and a row wire as described in connection with FIG. 1 , respectively.

[0055] Cross-point device 200 may include an RRAM device 201 and a transistor 203.

A transistor is a three-terminal device, which may be marked as gate (G), source (S), and drain (D), respectively. The transistor 203 may be serially connected to RRAM device 201. As shown in FIG. 2, the first electrode of the RRAM device 201 may be connected to the drain of transistor 203. The second electrode of the RRAM device 201 may be connected to the bitline 211. The source of the transistor 203 may be connected to the wordline 215. The gate of the transistor 203 may be connected to the select line 213. RRAM device 201 may include one or more RRAM

10 devices as described in connection with FIGS. 3A-7 below. Cross-point device 200 may also be referred to as in a 1-transitor-l -resistor (1T1R) configuration. The transistor 203 may perform as a selector as well as a current controller, which may set the current compliance, to the RRAM device 201 during programing. The gate voltage on transistor 203 can set current compliances to cross-point device 200 during programming and can thus control the conductance and analog behavior of cross-point device 200. For example, when cross-point device 200 is set from a high-resistance state to a low-resistance state, a set signal (e.g., a voltage signal, a current signal) may be provided via the bitline (BL) 211. Another voltage, also referred as a select voltage or gate voltage, may be applied via the select line (SEL) 213 to the transistor gate to open the gate and set the current compliance, while the wordline (WL) 215 may be set to ground. When cross- point device 200 is reset from the low-resistance state to the high-resistance state, a gate voltage may be applied to the gate of the transistor 203 via the select line 213 to open the transistor gate. Meanwhile, a reset signal may be sent to the RRAM device 201 via the wordline 215, while the bitline 211 may be set to ground.

[0056] FIGS. 3A, 3B, and 3C illustrate cross-sectional views of example RRAM devices in accordance with some embodiments of the present disclosure. RRAM devices 300a, 300b, and 300c may correspond to an RRAM device in an initial state, a low-resistance state, and a high-resistance state, respectively.

[0057] As shown in FIG. 3 A, RRAM device 300a may include a substrate 310, a first electrode 320, a switching oxide layer 330, and a second electrode 340. RRAM device 300a may further include one or more other components for implementing in-memory computing applications.

[0058] Substrate 310 may include one or more layers of any suitable material that may serve as a substrate for an RRAM device, such as silicon (Si), silicon dioxide (S1O2), silicon nitride (S13N4), aluminum oxide (AI2O3), aluminum nitride (AIN), etc. In some embodiments, substrate 310 may include diodes, transistors, interconnects, integrated circuits, etc. In some embodiments, the substrate may include a driving circuit including one or more electrical circuits (e.g., an array of electrical circuits) that may be individually controllable. In some embodiments, the driving circuit may include one or more complementary metal -oxide- semi conductor (CMOS) drivers.

[0059] First electrode 320 may be and/or include any suitable material that is electronically conductive and non-reactive to the switching oxide. For example, first electrode

11 320 may include platinum (Pt), palladium (Pd), iridium (Ir), titanium nitride (TiN), tantalum nitride (TaN), etc.

[0060] Switching oxide layer 330 may include one or more transition metal oxides, such as TaO _x , HfO _x , TiO _x , NbO _x , ZrO _x , etc., in binary oxides, ternary oxides, and high order oxides.

In some embodiments, the chemical stability of the non-reactive material in first electrode 320 may be higher than that of the transition metal oxide(s) in switching oxide layer 330.

[0061] Second electrode 340 may include any suitable metallic material that are electronically conductive and reactive to the switching oxide. For example, the metallic material in second electrode 340 may include Ta, Hf, Ti, TiN, TaN, etc. Second electrode 340 may be reactive to the switching oxide and may have suitable oxygen solubility to adsorb some oxygen from the switching oxide layer 330 and create oxygen vacancies in the switching oxide layer 330. In other words, the reactive metallic material(s) in second electrode 340 may have suitable oxygen solubility and/or oxygen mobility. In some embodiments, second electrode 340 not only may be able to create oxygen vacancies in switching oxide layer 330 (e.g., by scavenging oxygen), but also may function as oxygen reservoir or source to the switching oxide layer 330 during cell programming.

[0062] RRAM device 300a may have an initial resistance (also referred to herein as the

“virgin resistance”) after it is fabricated. The initial resistance of RRAM device 300a may be changed and RRAM device 300a may be switched to a state of a lower resistance via a forming process. For example, a suitable voltage or current may be applied to RRAM device 300a. The application of the voltage to RRAM device 300a may induce the metallic material(s) in the second electrode to absorb oxygen from the switching oxide layer 330 and create oxygen vacancies in the switching oxide layer 330. As a result, a conductive channel (e.g., a filament) which is oxygen vacancy rich may form in the switching oxide layer 330. For example, as illustrated in FIG. 3B, a conductive channel 335a may be formed in the switching oxide layer 330. As shown, conductive channel 335a may be formed from the second electrode 340 to the first electrode 320 across the switching oxide layer 330. RRAM device 300b may be regarded as being operated in a filamentary mode where the electric conduction is dominated by the conduction via the metallic filament while the electric conduction by the electronic defects in the switching oxide may be negligible. RRAM device 300b may be reset to a high-resistance state. For example, a reset signal (e.g., a voltage signal or a current signal) may be applied to RRAM device 300b during a reset process. In some embodiments, the set signal and the reset signal may have opposite polarity, i.e., a positive signal and a negative signal, respectively. The

12 application of the reset signal may cause oxygen to drift back to the switching oxide layer 330 and recombine with one or more of the oxygen vacancies. For example, an interrupted conductive channel 335b as shown in FIG. 3C may be formed in the switching oxide layer 330 during the reset process. As shown, the conductive channel may be interrupted with an oxide gap between the interrupted conductive channel 335b and the first electrode 320. The lateral dimension of conductive channel 335b may be smaller than that of the conductive channel 335a. In some embodiments, conductive channel 335b does not continuously connect the first electrode 320 and the second electrode 340. An oxide gap 335c is located between the interrupted filament 335b and the first electrode 320. RRAM device 300c may be regarded as being operated in a non-filamentary mode where electric conduction may be dominated by electronic defects in the switching oxide gap 335c that may have a much higher resistance than that of the interrupted filament 335b. RRAM device 300a-c may be electrically switched between the high- resistance state and the low-resistance state by applying suitable programming signals (e.g., voltage signals, current signals, etc.) to the RRAM device.

[0063] As described above, it may be necessary to operate the RRAM device in a non- filamentary mode to achieve a desired high resistance (e.g., a resistance higher than Go). For example, as shown in FIG. 3C, an electric path including an interrupted filament 335b and an oxide gap 335c located between the interrupted filament 335b and the first electrode is formed. Accordingly, defect engineering and controlled electronic defects in the switching oxide are important for applications requiring high resistance RRAM with resistances higher than Go. [0064] FIGS. 4A-4F are schematic diagrams illustrating cross-sectional views of example structures 400a, 400b, 400c, 400d, 400e, and 400f of RRAM devices in accordance with some embodiments of the present disclosure.

[0065] As illustrated FIG. 4A, a first electrode 420 may be fabricated on a substrate 410.

The first electrode 420 and the substrate 410 may correspond to the first electrode 320 and the substrate 310 as described in conjunction with FIGS. 3A, 3B, and 3C, respectively.

[0066] As illustrated FIG. 4B, an interface layer 422 may be fabricated on the first electrode 420. The interface layer 422 (also referred to herein as the “first interface layer”) may be and/or include a discontinuous film 422a. For example, the discontinuous film 422a may include one or more pores 424. The pore(s) 424 (also referred to herein as the “first pores”) may have any suitable size and/or dimension and may be dispersed randomly on the interface layer 422. While certain number of pores are illustrated in FIG. 4B, this is merely illustrative. The discontinuous film 422a may include any suitable number of pores. In some embodiments, a

13 thickness of the interface layer 422 and/or the discontinuous film 422a may be between about 0.2nm and about 0.5nm. In some embodiments, the discontinuous film 422a may be an AI2O3 film having a thickness equal to or less than 0.5nm. In some embodiments, the discontinuous film 422a may be and/or include an AI2O3 film having a thickness less than lnm.

[0067] As illustrated in FIG. 4C, a switching oxide layer 430 may be fabricated on the interface layer 422. The switching oxide layer 430 may include one or more transition metal oxides, such as TaO _x , HfO _x , TiO _x , NbO _x , ZrO _x , etc., in binary oxides, ternary oxides, and high order oxides, wherein x may be used to indicate the oxide being oxygen deficient compared to its full (or terminal) oxide and the value of x may be varied from the oxygen to metal atomic ratio in the stoichiometry of its full oxide, such as x<2.0 for HfO _x (where HfCk being the full oxide), and x<2.5 for TaO _x (where Ta ₂ 0s being the full oxide). As an example, the switching oxide layer 430 may include Ta ₂ 0s. As the other example, the switching oxide layer 430 may include HfCk. [0068] In some embodiments, during the fabrication of the switching oxide layer 430, one or more portions of the transition metal oxides may be disposed on the first electrode 420 through one or more pores 424. As such, the switching oxide layer 430 may contact one or more portions of the first electrode 420.

[0069] In some embodiments, the interface layer 422 may contain a first material that is more chemically stable than the transition metal oxide(s) in the switching oxide layer 430. As a result, the first material may not react with the transition metal oxide(s) of the switching oxide layer 430. As an example, the switching oxide of the switching oxide layer may be and/or include one or more transition metal oxides, such as at least one of HfO _x or TaO _y , wherein x<2.0, and wherein y<2.5, and the first material may include AI2O3, MgO, Y2O3, La2C>3, etc.

[0070] Referring to FIG. 8, materials that are more chemically stable than the transition metal oxide(s) in the switching oxide layer 430 may be identified using Ellingham diagrams. As illustrated, an Ellingham diagram 800 plots the Gibbs free energy change for an oxidation reaction as a function of temperature. The chemical stability of the materials may be determined based on the Gibbs formation energy values of the materials. The Gibbs free energy shown on the vertical axis in FIG. 8 represents the free energy of formation of an oxide (containing 1 mole of oxygen) and the horizontal axis represents the temperature in Kelvin. As shown, the relative stability of these oxides increases from Ta20s, T1O2, Hf02to AI2O3. AI2O3 is more stable than Hf0 ₂ at room temperature. It should be noted that A1 metal melts at 933K (660°C) and A1 liquid has higher entropy than A1 solid. AI2O3 becomes less stable than Hf0 ₂ at higher temperatures. In an Ellingham diagram, the curve of the material of the interface layer 422 may be below the

14 curve corresponding to the transition metal oxide(s) of the switching oxide layer 430. As an example, AI2O3 may be used as the first material in some embodiments in which the transition metal oxide in the switching oxide layer 430 contains HfC or Ta ₂ 0s. The first material containing AI2O3 does not react with the transition metal oxide containing HfC or Ta2(¾ during the setting process and resetting process.

[0071] As shown in FIG. 4D, a second electrode 440 may be fabricated on the switching oxide layer 430. The second electrode 440 may function as a defect engineering layer for generating defects in the switching oxide layer 430. In some embodiments, the second electrode 440 may include one or more top electrodes 700 as described in connection with FIG. 7. In some embodiments, the second electrode 440 fabricated on the switching oxide layer 430 may include one or more alloys. Each of the alloys may contain two or more metallic elements. Each of the alloys may include a binary alloy (e.g., an alloy containing two metallic elements), a ternary alloy (e.g., an alloy containing three metallic elements), a quaternary alloy (e.g., an alloy containing four metallic elements), a quinary alloy (e.g., an alloy containing five metallic elements), a senary alloy (e.g., an alloy containing six metallic elements), and/or a high order alloy (e.g., an alloy containing more than six metallic elements). In some embodiments, the second electrode 440 may include one or more alloys containing a first metallic element and one or more second metallic elements. Each of the second metallic elements may be less or more reactive to the transition metal oxide in the switching oxide layer than the first metallic element. In some embodiments, the first metallic element may be Ta. The second metallic elements may include one or more of W, Hf, Mo, Nb, Zr, etc. In some embodiments, the ratio of the first metallic element to the second metallic element(s) in an alloy in the second electrode 340 may be about 50 atomic percent. In some embodiments, the suitable ratio of the first metallic element to the second metallic element in the alloy may be optimized from the entire composition range. During a forming process, the second metallic element(s) may create fewer oxygen vacancies in the switching oxide layer than the first metallic element. As such, the lateral size of the filament formed in an RRAM device comprising a second electrode containing the alloy may be smaller than that of the filament formed in an RRAM device comprising a second electrode made of only the first metal. The implementation of alloy containing Ta in the second electrode 440 in the RRAM device may result in a less abrupt forming process, reduce the forming voltage, reduce the reset current, and reduce voltage and/or current requirements in subsequent operation processes. Furthermore, incorporating the alloy containing Ta in the second electrode 440 may also generate suitable electronic defects in the switching oxides for performing operations that

15 require RRAMs with high resistance. Incorporating Ta in the second electrode 440 can also generate defects in the switching oxides. Those defects are commonly regarded as oxygen vacancies which are structural defects that may enhance the migration of oxygen ions in the switching oxide layer. Oxygen vacancies as electronic defects do not affect operations of the RRAM device in a filamentary mode since the electric conduction in the filamentary mode is dominated by the metallic filament, not by the electronic defects in the switching oxide layer. However, in the non-filamentary mode, these oxygen vacancies may perform as electronic defects (to trap and de-trap electrons) instead of as structural defects (to migrate oxygen ions).

An alloy incorporating Ta may provide electronic defects in the switching oxide layer 335c for electric conduction with high resistance in the non-filamentary mode. The RRAM device with the electronic defects may present certain electronic behaviors that are critical to IMC applications, such as low current, analog resistance, multilevel resistance, I-V (current-voltage) linearity, etc.

[0072] As an example, the second electrode 440 may include one or more alloys containing Ta (also referred to as “Ta alloys”). Each of the Ta alloys may include Ta and one or more other metallic elements (e.g., Hf, W, Mo, Nb, Zr, etc.). As an example, the second electrode 440 may include one or more binary alloys containing Ta. Examples of the binary alloys containing Ta include a Ta-Hf alloy, a Ta-W alloy, a Ta-Mo alloy, a Ta-Nb alloy, a Ta-Zr alloy, etc. As another example, the second electrode 440 may include one or more ternary alloys containing Ta. Examples of the ternary alloys containing Ta include a Ta-Hf-Mo alloy, a Ta-Hf- Nb alloy, a Ta-Hf-W alloy, a Ta-Hf-Zr alloy, a Ta-Mo-Nb alloy, a Ta-Mo-W alloy, a Ta-Mo-Zr alloy, a Ta-Nb-W alloy, a Ta-Nb-Zr alloy, a Ta-W-Zr alloy, etc. As still another example, the second electrode 440 may include one or more quaternary alloys containing Ta. Examples of the quaternary alloys containing Ta include a Ta-Hf-Mo-Nb alloy, a Ta-Hf-Mo-W alloy, a Ta-Hf- Mo-Zr alloy, a Ta-Hf-Nb-W alloy, a Ta-Hf-Nb-Zr alloy, a Ta-Mo-Nb-W alloy, a Ta-Mo-Nb-Zr alloy, a Ta-Nb-W-Zr alloy, etc. As a further example, the second electrode 440 may include one or more quinary alloys containing Ta. Examples of the quinary alloys containing Ta include a Ta-Hf-Mo-Nb-W alloy, a Ta-Mo-Nb-W-Zr alloy, a Ta-Hf-Nb-W-Zr alloy, a Ta-Hf-Mo-W-Zr alloy, a Ta-Hf-Mo-Nb-Zr alloy, etc. As still a further example, the second electrode 440 may include a senary alloy containing Ta, such as a Ta-Hf- Mo-Nb-W-Zr alloy. As a further example, the second electrode 440 may include a high order alloy containing Ta. In some embodiments, the high order alloy may further contain vanadium (V).

16 [0073] In some embodiments, the second electrode 440 may include a plurality of alloys.

Each of the alloys may be a Ta alloy containing Ta and one or more other metallic elements (e.g., Hf, W, Mo, Nb, Zr, etc.). The Ta alloy may be and/or include a binary alloy, a ternary alloy, a quaternary alloy, a quinary alloy, a senary alloy, a high-order alloy, etc. As an example, the second electrode 440 may include two or more of a first alloy containing Ta, a second alloy containing Ta, a third alloy containing Ta, a fourth alloy containing Ta, a fifth alloy containing Ta, and a sixth alloy containing Ta. In some embodiments, the first alloy containing Ta, the second alloy containing Ta, the third alloy containing Ta, the fourth alloy containing Ta, the fifth alloy containing Ta, and the sixth alloy containing Ta may be a binary alloy, a ternary alloy, a quaternary alloy, a quinary alloy, a senary alloy, and a high-order alloy, respectively.

[0074] In some embodiments, multiple alloys in the second electrode 440 may correspond to combinations of the same number of metallic elements. For example, the first alloy containing Ta and the second alloy containing Ta may include a first binary alloy containing Ta (e.g., a Ta-W alloy) and a second binary alloy containing Ta (e.g., a Ta-Mo alloy), respectively. As another example, the first alloy containing Ta and the second alloy containing Ta may include a first ternary alloy containing Ta (e.g., a Ta-Hf-Mo alloy) and a second ternary alloy containing Ta (e.g., a Ta-Hf-Nb alloy), respectively. As a further example, the second electrode 440 may include a plurality of alloy systems. Each of the alloy systems may contain the alloys containing the mixtures of certain metallic elements with varying compositions. For example, a binary system may include one or more binary alloys of two metallic elements (e.g., Ta and Hf) with varying compositions. Each of the binary alloys may be a combination of the two metallic elements with a certain composition. As another example, a ternary system may include one or more ternary alloys of three metallic elements (e.g., Ta, Hf, and W) with varying compositions. Each of the ternary alloys may be a combination of the three metallic elements with a certain composition. In some embodiments, the second electrode 440 may include a Ta alloy system containing one or more alloy systems. In some embodiments, the Ta alloy system may include two or more alloy systems. For example, the Ta alloy system may contain a senary system containing one or more alloys of Ta, Hf, W, Mo, Nb, and Zr. The Ta alloy system may further include one or more binary systems, ternary systems, quaternary systems, and/or quinary systems containing Ta alloys. The binary systems may include one or more of a Ta-Hf alloy system, a Ta-W alloy system, a Ta-Mo alloy system, a Ta-Nb alloy system, and/or a Ta-Zr alloy system. The ternary systems may include one or more of a Ta-Hf-Mo alloy system, a Ta-Hf-Nb alloy system, a Ta-Hf-W alloy system, a Ta-Hf-Zr alloy system, a Ta-Mo-Nb alloy system, a Ta-

17 Mo-W alloy system, a Ta-Mo-Zr alloy system, a Ta-Nb-W alloy system, a Ta-Nb-Zr alloy system, and/or a Ta-W-Zr alloy system. The quaternary systems may include one or more of a Ta-Hf-Mo-Nb alloy system, a Ta-Hf-Mo-W alloy system, a Ta-Hf-Mo-Zr alloy system, a Ta-Hf- Nb-W alloy system, a Ta-Hf-Nb-Zr alloy system, a Ta-Mo-Nb-W alloy system, a Ta-Mo-Nb-Zr alloy system, a Ta-Nb-W-Zr alloy system. The quinary systems may include one or more of a Ta-Hf-Mo-Nb-W alloy system, a Ta-Mo-Nb-W-Zr alloy system, a Ta-Hf-Nb-W-Zr alloy system, a Ta-Hf-Mo-W-Zr alloy system, and/or a Ta-Hf-Mo-Nb-Zr alloy system. Each of the alloy systems contained in the second electrode 440 may have unique thermodynamic and kinetic characteristics and may be regarded as an electrode component. As such, the second electrode 440 may include multiple electrode components for providing multiple state variables that may lead to rich dynamics with various time constants for computing and learning. For example, each electrode component may have different reactivity to the switching oxide or the affinity for oxygen. Each electrode element may have different diffusivity (e.g., self-diffusion, inter diffusion, diffusion time constants, etc.). The second electrode 440 with multiple components may provide the multiple dynamic behavior for IMC applications. As such, the RRAM device incorporating the multi-component second electrode may present dynamic memri stive behavior in multiple dimensions. Furthermore, the implementation of multi-component second electrode may also generate suitable electronic defects in the switching oxides for performing operations that require RRAMs with high resistance. The RRAM device with the electronic defects may present certain electronic behaviors that are critical to IMC applications, such as low current, analog resistance, multilevel resistance, I-V (current-voltage) linearity, etc.

[0075] FIGS. 4E and 4F illustrate semiconductor devices 400e and 400f that may correspond to a low-resistance state and a high-resistance state of the RRAM device 400d, respectively. The incorporation of the discontinuous film 422 may reduce the contact area between the switching oxide layer 430 and the first electrode 420. As illustrated in FIG. 4E, a conductive channel 435a (e.g., a filament) may be formed from the second electrode 440 to the first electrode 420 through the interface layer 422 and the switching oxide layer 430. As illustrated in FIG. 4F, an interrupted conductive channel 435b may be formed in the switching oxide layer 430 during the reset process. Compared with FIGS.3B and 3C, the lateral size of the conductive channel 435a and the interrupted conductive channel 435b may be smaller than that of 335a and 335b respectively. As such, the lateral size of the filament formed in an RRAM device with discontinuous film 422 may be smaller than that of the filament formed in an RRAM device without a porous and/or discontinuous film formed between the first electrode and the

18 switching oxide layer. For example, forming an RRAM device with a non-porous or continuous film (e.g., with a non-porous or continuous interface layer) fabricated in a virgin state may need an electric breakdown of the non-porous or continuous film to establish an electric path passing through the RRAM device. The electric breakdown of the non-porous or continuous film may be abrupt and may over form or over grow the conductive filament in the RRAM device, which may deteriorate or prevent the subsequent low current operations of the RRAM device. The incorporation of the discontinuous film 422 into the RRAM device may result in a less abrupt forming process, reduce the forming voltage, reduce the reset current, and reduce voltage and/or current requirements in subsequent operation processes.

[0076] In some embodiments, an RRAM device may include multiple interface layers fabricated between the first electrode and the second electrode. Each of the interface layers may include a discontinuous film as described in connection with FIG. 4B. For example, as illustrated in FIG 5A, a semiconductor device 500a may be fabricated, by fabricating an interface layer 532 (also referred to as the “second interface layer”) on the semiconductor structure 400c as described in connection with FIG. 4C. In some embodiments, the second interface layer 532 may include a discontinuous film 532a of a second material. The second material may be more chemically stable than the at least one transition metal oxide in the switching oxide layer 430.

As an example, the second material may include AI2O3, MgO, Y2O3, La ₂ 0 ₃ , etc. The second material may or may be the same as the first material.

[0077] The discontinuous film 532a may include one or more pores 534 (also referred to as the “one or more second pores”). The pore(s) 534 may have any suitable size and/or dimension. Multiple pores 534 may or may not have the same size and/or dimension. In some embodiments, the second interface layer 532 and/or the second discontinuous film 532a may include multiple pores 534 dispersed randomly on the second discontinuous film 532. The discontinuous film 532a may include any suitable number of pores.

[0078] In some embodiments, a thickness of the second interface layer 532 and/or the second discontinuous film (also referred to as the “second thickness”) may be between about 0.2 nm and about 0.5 nm. As another example, the second interface layer 532 may include a discontinuous AI2O3 film having a thickness equal to or less than 0.5 nm. In some embodiments, the second interface layer 532 may include a discontinuous AI2O3 film having a thickness less than lnm. The second thickness of the second interface layer 532 may or may not be the same as the first thickness of the first interface layer 422.

19 [0079] As illustrated in FIG. 5B, a second electrode 540 may be fabricated on the second interface layer 532 to fabricate a semiconductor device 500b. The second interface layer 532 may thus be positioned between the switching oxide layer 430 and the second electrode 540.

The second electrode 540 may function as a defect engineering layer for generating electronic defects in the switching oxide layer 430. The second electrode 540 may be and/or include the second electrode 440 as described in conjunction with FIGS. 4D-4F. In some embodiments, during the fabrication of the switching oxide layer 430, one or more portions of the second electrode 540 may be disposed on the switching oxide layer 430 through one or more pores 534. As such, the second electrode 540 may contact one or more portions of the switching oxide layer 430 through the one or more pores 534.

[0080] FIGS. 5C and 5D illustrate semiconductor devices 500c and 500d that may correspond to a low-resistance state and a high-resistance state of the RRAM device 500b, respectively. The incorporation of both the first interface layer 422 and the second interface layer 532 may further reduce the contact area between the switching oxide layer 430 and the first electrode 420 and the contact area between the switching oxide layer 430 and the second electrode 540. As illustrated in FIG. 5C, a conductive channel 535a (e.g., a filament) may be formed from the first electrode 420 to the second electrode 540 through the interface layer 422, the switching oxide layer 430, and the second discontinuous film 532. As illustrated in FIG. 5D, an interrupted conductive channel 535b may be formed in the switching oxide layer 530 during the reset process. Compared with FIGS. 3B and 3C, the lateral size of the conductive channel 535a and the interrupted conductive channel 535b may be further smaller than that of 335a and 335b, respectively. As such, the lateral size of the filament formed in an RRAM device with both the interface layer 422 and the second interface layer 532 may be further reduced, resulting in a less abrupt forming process, reducing the forming voltage, reducing the reset current, and reducing voltage and/or current requirements in subsequent operation processes.

[0081] In some embodiments, an interface layer may be fabricated on a switching oxide layer of an RRAM device in accordance with some embodiments of the present disclosure. For example, as illustrated in FIG 6A, a switching oxide layer 630 may be fabricated on the semiconductor device 400a as described in FIG. 4A. An interface layer 632 may be fabricated on the switching oxide layer 630 to fabricate a semiconductor device 600a. The switching oxide layer 630 may be and/or include the switching oxide layer 430 as described in conjunction with FIGS. 4C-4F. In some embodiments, the interface layer 632 may include a discontinuous film of a third material. The third material may be more chemically stable than the at least one transition

20 metal oxide in the switching oxide layer 630. As an example, the third material may include AI2O3, MgO, Y2O3, La2(¾, etc.

[0082] The discontinuous film 632a may include one or more pores 634 (also referred to herein as the “one or more third pores”). The pore(s) 634 may have any suitable size and/or dimension and may be dispersed randomly on the interface layer 632. In some embodiments, a thickness of the interface layer 632 (also referred to as the “third thickness”) may be between about 0.2 nm and about 0.5 nm. As another, the interface layer 632 may have a thickness equal to or less than 0.5 nm thickness. As a further example, the interface layer 632 may have a thickness less than 1 nm.

[0083] As illustrated in FIG. 6B, a second electrode 640 may be fabricated on the interface layer 632 to fabricate a semiconductor device 600b. The second electrode 640 may function as a defect engineering layer for generating electronic defects in the switching oxide layer 630. The interface layer 632 may thus be positioned between the switching oxide layer 630 and the second electrode 640. The second electrode 640 may be and/or include the second electrode 440 as described in conjunction with FIGS. 4D-4F. In some embodiments, during the fabrication of the second electrode 640, one or more portions of the second electrode 640 may be disposed on the switching oxide layer 630 through one or more pores 634. As such, the second electrode 640 may contact one or more portions of the switching oxide layer 630.

[0084] FIGS. 6C and 6D illustrate semiconductor devices 600c and 600d that may correspond to a low-resistance state and a high-resistance state of the RRAM device 600b, respectively. The incorporation of the discontinuous film 632 may reduce the contact area between the switching oxide layer 630 and the second electrode 640. As illustrated in FIG. 6C, a conductive channel 635a (e.g., a filament) may be formed from the second electrode 640 to the first electrode 420 through the switching oxide layer 630 and the discontinuous film 632. As illustrated in FIG. 6D, an interrupted conductive channel 635b may be formed in the switching oxide layer 630 during the reset process. Compared with FIGS. 3B and 3C, the lateral size of the conductive channel 635a and the interrupted conductive channel 635b may be smaller than that of 335a and 335b respectively. As such, the lateral size of the filament formed in an RRAM device with a discontinuous film 632 may be smaller than that of the filament formed in an RRAM device without the discontinuous film 632. The incorporation of a discontinuous film 632 into the RRAM device may result in a less abrupt forming process, reduce the forming voltage, reduce the reset current, and reduce voltage and/or current requirements in subsequent operation processes. Incorporating the discontinuous film 632 in the RRAM device may assist in

21 electronic defect engineering as it may increase the device resistance and control the number of electronic defects that participate in high resistance operations. More particularly, electronic defects alone may reduce device resistance or make the device leaky. Incorporating the porous and/or discontinuous interface layer may increase device resistance by restricting an electric path through the switching oxide. Accordingly, a combination of porous/discontinuous interface films 422 or 532 and the second electrode 540 may generate and control suitable electronic defects for applications requiring high resistance RRAMs.

[0085] FIG. 7 is a schematic diagram illustrating cross-sectional views of an example

700 of a top electrode in accordance with some embodiments of the present disclosure.

[0086] The top electrode 700 may function as a defect engineering layer as described herein. As shown, the top electrode 700 may include a first layer 710 and a second layer 720. The first layer 710 may include a first metallic material that may scavenge oxygen from the transition metal oxide(s) of the switching oxide layer. The second layer 720 may include a second metallic material that may scavenge oxygen from the transition metal oxide(s) of the switching oxide layer. The first metallic material may have oxygen solubilities and may react with and scavenge oxygen from the transition metal oxide(s) of the switching oxide layer. The oxide of the first metallic material may have less chemical stability than the first material of the first discontinuous film and the second material of the second discontinuous film. As a result, the first metallic material may not chemically reduce the first discontinuous film and the second discontinuous film. As described above, the Ellingham diagrams may be employed to determine the comparative chemical stability of two or more elements.

[0087] The first metallic material and the second metallic material may include different chemical elements and may have different affinities for oxygen and/or different thermodynamic and kinetic properties. The first metallic material and the second metallic material may be immiscible. In some embodiments, the first metallic material may include Ti. The second metallic material may include Ta. In some embodiments, the first layer 710 may be and/or include a layer of Ti metal (e.g., a Ti film). The second layer 720 may be and/or include a layer of Ta metal (e.g., a Ta film). As shown in the Ta-Ti binary phase diagram of FIG. 9, Ta and Ti phases are immiscible and have minimum mutual solubilities at 300K (27°C) which is around the operating temperatures of RRAM devices (e.g., temperatures around, below, or above room temperature). As such, the addition of Ti into the RRAM device may not affect the operation mechanism of the Ta filament in the switching oxide and the switching mechanism of RRAM devices as described herein. The RRAM devices as described in FIG. 7 may thus be used for

22 IMC applications that require RRAM devices with excellent performance in analog behaviors, linearity, retention, reliability, etc. The immiscibility and minimum mutual solubilities between Ta and Ti phases may also enable a thermodynamic equilibrium between the second layer 720 and the first layer 710. As a result, a thin Ti film can function as designed without reacting with the Ta film or being dissolved by the Ta film.

[0088] Furthermore, Ti may readily scavenge oxygen from the switching oxide because it has higher affinity for oxygen than Ta. As such, the incorporation of the first layer 710 into the RRAM device may further improve the performance of the RRAM device by reducing the forming voltage required in the RRAM forming process and the current and voltage requirements in subsequent operations. For example, the second electrode 440 in FIGS. 4D-4F, the second electrode 540 in FIGS.5B-5D, and/or the second electrode 640 in FIGS. 6B-6D as described above may be and/or include the second electrode 700. During a forming process, both the first metallic material and the second metallic material may generate oxygen vacancies in the switching oxide layer 330, 430, and/or 630. Compared with FIGS. 3B and 3C, the lateral size of the conductive channel and the interrupted conductive channel may be smaller than that of 335a and 335b, respectively. As Ti has a higher affinity for oxygen than Ta, the virgin resistance of an RRAM device including the top electrode 700 may be lower than that of an RRAM device that does not include the top electrode 700, resulting in the lower forming voltage, the lower reset voltage, the low reset current, etc.

[0089] Ti may also readily store oxygen during a set process (when oxygen is migrating from switching oxide to the second electrode). This may enable the second electrode to store oxygen during the reset process and may thus prevent device failures (which can be caused by the presence of oxygen molecules between switching oxide and the second electrode) and/or operation failures (which can be caused by the oxygen migrating back to the switching oxide once the set voltage being removed, or the switch being volatile). Furthermore, the incorporation of a Ti film into the RRAM device may generate electronic defects in the switching oxide of the RRAM device. The electronic defects may be electrons trapped in oxygen vacancies. Under an electric field, the trapped electrons may hop from one trap site to another trap site or may be excited to the conduction band with a lower excitation energy, generating an electron flow when the RRAM device is in a high resistance state.

[0090] The first layer 710 may be grown to a suitable thickness so that the first metallic material (e.g., Ti) of the first layer 710 may function as described above without affecting the formation of the filament comprising the second metallic material (e.g., a Ta filament) in the

23 switching oxide layer 330. In some embodiments, a thickness of the first layer 710 may be between about 0.2nm and about 5nm. In some embodiments, a thickness of the first layer 710 may be between about 0.5nm and about 2nm. In some embodiments, a thickness of the first layer 710 may be about lnm. In some embodiments, a thickness of the first layer 710 may be less than lnm. The second layer 720 may be thicker than the first layer 710. In some embodiments, a thickness of the second layer 720 may be between 5nm and 300nm. For example, the thickness of the second layer 720 may be between about lOnm and about lOOnm.

In some embodiments, the thickness of the second layer 720 may be between about lOnm and about 200nm. In some embodiments, the thickness of the second layer 720 may be about 50nm. A thickness of the first electrode may be between about 5nm and lOOnm. In some embodiments, the thickness of the first electrode may be about 30nm. In some embodiments, a dimension (e.g., a critical dimension) of RRAM device 400d, 500b, and/or 600b may be between 1 m m and single digit nanometers. In some embodiments, the critical dimension of RRAM device 400d, 500b, and/or 600b may be about or less than 0.28 pm, and/or between lpm and 1 nanometer. In some embodiments, the critical dimension of RRAM device 400d, 500b, and/or 600b may be between lpm and 2nm. In some embodiments, the critical dimension of RRAM device 400d, 500b, and/or 600b may be between lpm and 5nm. In some embodiments, the critical dimension of RRAM device 400d, 500b, and/or 600b may be at the single-digit nanoscale (e.g., between about lnm and about 9nm).

[0091] In one implementation, the second layer 720 may be fabricated directly on the

First layer 710. For example, as shown in FIG. 7, a surface of the second layer 720 may directly contact one or more portions of a surface of the first layer 710. In another implementation, one or more other layers of suitable materials may be deposited between the first layer 710 and the second layer 720.

[0092] When functioning as an insulator, the transition metal oxide in the switching oxide layer described herein (e.g., switching oxide layer 430, 530, and/or 630) may have a band gap between its valance band and conduction band. An energy higher than the energy represented by the band gap is required to excite an electron from the valance band to the conduction band where the electron can participate in conduction. It is important to have proper electronic defects with suitable densities as electron traps within the band gap for a trapped electron being excited to the conduction band (thermal emission) or hopping (tunneling) from one trap site to the other trap site. The incorporation of the defect engineering layer(s) as described herein (e.g., an alloy containing Ta 440, the first layer 710, etc.) may produce

24 electronic defects with suitable densities in the transition metal oxide(s) in the switching oxide layer 430, 530, and/or 630. The electronic defects may include, for example, oxygen vacancy defects in the transition metal oxide(s) (e.g., sub-stoichiometric oxides of TaO _x or HfO _x . The electronic defects may can assist charges in transporting through the transition metal oxide(s) at room temperature and above. In some embodiments, the charges may be ionic, such as oxygen ion O ² that carries -2 charges or oxygen vacancy Vo ⁺² that carries +2 charges. In some embodiments, the charges may be electronic, such as an electron e ¹ that carries a -1 charge. In this case, under an electric field, the trapped electrons may be excited to conduction band or hopping from one trap site to another trap site to generate an electron flow when the RRAM device incorporating the defect engineering device is operated in a non-filamentary state.

[0093] The incorporation of the first layer 710 in the RRAM device may thus reduce the virgin resistance of the RRAM device, reduce the forming voltage, and reduce the reset current, resulting in a less abrupt forming process, a filament with lower conductance, lower voltage and lower current in the subsequent operation process, and produce suitable electronic defects in switching oxides for IMC operations in high resistance states.

[0094] While certain components of RRAM devices 400d-f, 500b-d, and 600b-d are shown in FIGS. 4D-4F, 5B-5D, and 6B-6D, this is merely illustrative. RRAM devices 400d-f, 500b-d, and 600b-d may include one or more other layers of suitable materials for implementing IMC applications. For example, one or more interfacial layers (not shown) may be fabricated between the switching oxide layer and one or more of the second electrode and the first electrode to improve the interface stability and device performance.

[0095] FIG. 10 is a flow diagram illustrating an example 1000 of a method for fabricating an RRAM device according to some embodiments of the disclosure.

[0096] At block 1010, a first electrode may be fabricated on a substrate. Fabricating the first electrode may involve depositing one or more layers of one or more nonactive metals, such as Pt, Pd, Ir, etc. utilizing a physical vapor deposition (PVD) technique, a chemical vapor deposition (CVD) technique, a sputtering deposition technique, an atomic layer deposition (ALD) technique, and/or any other suitable deposition technique. In some embodiments, fabricating the first electrode may involve depositing one or more layers of Pt. The first electrode may be and/or include first electrode 320, 420 as described in connection with FIGS. 3A-6D above.

[0097] At block 1020, an interface layer may be fabricated on the first electrode.

Fabricating the first interface layer may involve depositing a first material on the first electrode

25 to form a first discontinuous film of the first material. The first discontinuous film may contain one or more first pores. The first material may be more chemically stable than the transition metal oxide(s) in the switching oxide layer as described below. In some embodiments, the first material may include AI2O3, MgO, Y2O3, La2(¾, etc. The first interface layer may be and/or include the first interface layer 422 as described in connection with FIGS. 4B-5D above in some embodiments. In some embodiments, fabricating the first interface layer may involve depositing a layer of the first material having a suitable thickness to form the first discontinuous film. For example, fabricating the first interface layer may involve depositing the first material to a thickness between about 0.2nm and about lnm. The first discontinuous film may be deposited utilizing PVD, CVD, ALD, and/or any other suitable deposition technique.

[0098] At block 1030, a switching oxide layer may be fabricated on the interface layer.

The switching oxide layer may include one or more transition metal oxides. The transition metal oxides may include, for example, TaO _x , HfO _x , TiO _x , NbO _x , ZrO _x , etc. In some embodiments, during the fabrication of the switching oxide layer, one or more portions of the transition metal oxides may be disposed on the first electrode through one or more of the first pores. The switching oxide layer may be deposited utilizing PVD, CVD, ALD, and/or any other suitable deposition technique. The switching oxide layer may be and/or include switching oxide layer 430 as described in connection with FIGS. 4C-5D above.

[0099] At block 1040, a defect engineering layer may be fabricated on the switching oxide layer for generating electronic defects in the switching oxide layer. The defect engineering layer may function as a second electrode (e.g., the top electrode) of the RRAM device to be fabricated. The defect engineering layer may include one or more alloys. Each of the alloys may contain a first metallic element and one or more second metallic elements. Each of the second metallic elements and the first metallic element may have different reactivity to the transition metal oxide in the switching oxide layer. In some embodiments, the first metallic element may be Ta. The second metallic elements may be one or more of W, Hf, Mo, Nb, Zr, etc. Based on the binary phase diagrams involving Ta and the second metallic elements W, Hf, Mo, Nb, or Zr as shown in FIGS. 14A-14E, Ta-W (FIG. 14B), Ta-Mo (FIG. 14C), and Ta-Nb (FIG. 14D) form continuous solid solution, and Ta-Hf (FIG. 14A) and Ta-Zr (FIG. 14E) are immiscible at RRAM devices operation temperatures. No binary intermetallic compounds may form in all these binaries. No intermetallic compounds in these binary systems are advantageous for IMC applications involving an alloy electrode which may be readily fabricated and controlled. For example, fabricating the second electrode may involve fabricating the alloy by

26 co-sputtering the first metal and the second metal. As another example, fabricating the second electrode may involve sputtering from an alloy target (e.g., a Ta-W alloy, Ta-Hf alloy, Ta-Mo alloy, Ta-Nb alloy, Ta-Zr alloy, etc.) for a required composition between the first metallic element (e.g., a pure Ta metal) and the second metallic element(s) (e.g., a pure Hf metal).

[00100] In some embodiments, fabricating the defect engineering layer may involve fabricating multiple electrode components including Ta, Hf, Nb, Mo, W, and/or Zr. Each of the electrode components may be and/or include a binary alloy, a ternary alloy, a quaternary alloy, a quinary alloy, a senary alloy, and/or a high order alloy of Ta. For example, fabricating the defect engineering layer may involve fabricating the second electrode 340b with one or more alloys and/or alloy systems as described in connection with FIG. 4D. More particularly, for example, fabricating the second electrode may involve fabricating two or more of a first alloy containing Ta, a second alloy containing Ta, a third alloy containing Ta, a fourth alloy containing Ta, a fifth alloy containing Ta, and a sixth alloy containing Ta. The first alloy containing Ta, the second alloy containing Ta, the third alloy containing Ta, the fourth alloy containing Ta, the fifth alloy containing Ta, and the sixth alloy containing Ta may be a binary alloy, a ternary alloy, a quaternary alloy, a quinary alloy, a senary alloy, and a high-order alloy, respectively.

[00101] In some embodiments, fabricating the defect engineering layer may involve fabricating multiple layers of multiple metallic materials, such as layers 710 and 720 as described in conjunction with FIG. 7. In some embodiments, the defect engineering layer may be fabricated by performing one or more operations as described in connection with FIG. 13. [00102] The defect engineering layer may be fabricated utilizing PVD, CVD, AFD, and/or any other suitable deposition technique. The defect engineering layer may be and/or include second electrode 440 as described in connection with FIGS. 4D-4F above.

[00103] FIG. 11 is a flow diagram illustrating an example 1100 of a method for fabricating an RRAM device according to some embodiments of the disclosure.

[00104] At block 1110, a first electrode may be fabricated on a substrate. The first electrode may be fabricated on the substrate by performing one or more operations described in connection with block 1010 of FIG. 10. The first electrode may be and/or include first electrode 320 and/or 420 as described in connection with FIGS. 3A-6D above.

[00105] At block 1120, a first interface layer may be fabricated on the first electrode. Fabricating the first interface layer may involve fabricating a first discontinuous film of a first material. The first interface layer may be fabricated on the first electrode by performing one or more operation described in connection with block 1020 of FIG. 10. The first interface layer

27 may be and/or include the first interface layer 422 as described in connection with FIGS. 4B-5D above.

[00106] At block 1130, a switching oxide layer may be fabricated on the first interface layer. The switching oxide layer may be fabricated on the first interface layer by performing one or more operations described in connection with block 1030 of FIG. 10. The switching oxide layer may be and/or include switching oxide layer 430 as described in connection with FIGS. 4C-5D above.

[00107] At block 1140, a second interface layer may be fabricated on the switching oxide layer. Fabricating the second interface layer may involve fabricating a second discontinuous film of a second material that is more chemically stable than the transition metal oxide(s) of the switching oxide layer. In some embodiments, the second material may include AI2O3, MgO, Y2O3, La ₂ 0 ₃ , etc. In some embodiments, fabricating the second interface layer may involve depositing the second material to a suitable thickness (e.g., a thickness between 0.2nm and lnm) to form the second discontinuous film. The discontinuous film may be deposited utilizing PVD, CVD, ALD, and/or any other suitable deposition technique. The second interface layer may be and/or include second interface layer 532 as described in connection with FIGS. 5A-5D.

[00108] At block 1150, a defect engineering layer may be fabricated on the second interface layer for generating electronic defects in the switching oxide layer. The defect engineering layer may function as a second electrode (e.g., the top electrode) of the RRAM device to be fabricated. The defect engineering layer may be fabricated by performing one or more operations described in connection with block 1040 of FIG. 10. In some embodiments, during the fabrication of the defect engineering layer, one or more portions of the defect engineering layer may be deposited on the switching oxide layer through the one or more second pores. The defect engineering layer may be and/or include second electrode 540 as described in connection with FIGS. 5B-5D above.

[00109] FIG. 12 is a flow diagram illustrating an example 1200 of a method for fabricating an RRAM device according to some embodiments of the disclosure.

[00110] At block 1210, a first electrode may be fabricated on a substrate. The first electrode may be fabricated on the substrate by performing one or more operations described in connection with block 1010 of FIG. 10. The first electrode may be and/or include first electrode 320 as described in connection with FIGS. 3A-6D above.

[00111] At block 1220, a switching oxide layer may be fabricated on the first electrode. The switching oxide layer may include one or more transition metal oxides. The transition metal

28 oxides may include, for example, TaO _x , HfO _x , TiO _x , NbO _x , ZrO _x , etc. The switching oxide layer may be deposited utilizing PVD, CVD, ALD, and/or any other suitable deposition technique.

The switching oxide layer may be and/or include switching oxide layer 630 as described in connection with FIGS. 6A-6D above.

[00112] At block 1230, an interface layer may be fabricated on the switching oxide layer. The interface layer may include a discontinuous film of a material that is more chemically stable than the transition metal oxide(s) in the switching oxide layer. The interface layer may be fabricated on the switching oxide layer by performing one or more operations described in connection with block 1140 of FIG. 11. The interface layer may be and/or include the second interface layer 632 as described in connection with FIGS. 6A-6D above. As pointed out previously, the discontinuous or porous interface layer may contribute to electronic defect engineering in the switching oxide by controlling the number of electronic defects that may participate in the high resistance states operations.

[00113] At block 1240, a defect engineering layer may be fabricated on the interface layer. The defect engineering layer may function as a second electrode (e.g., the top electrode) of the RRAM device. The defect engineering layer may be fabricated on the interface layer by performing one or more operations described in connection with block 1040 of FIG. 10. The defect engineering layer may be and/or include the second electrode 640 as described in connection with FIGS. 6B-6D above.

[00114] FIG. 13 is a flow diagram illustrating an example 1300 of a method for fabricating a defect engineering layer of an RRAM device according to some embodiments of the disclosure. The defect engineering layer may function as a top electrode of the RRAM device.

[00115] At block 1310, a first layer of a first metallic material may be fabricated. The first metallic material may include a first metallic element, such as Ti, Hf, and Zr. The first layer of the first metallic material may be fabricated by depositing a first metal (e.g., Ti metal) utilizing PVD, CVD, sputtering, ALD, and/or any other suitable deposition technique. Fabricating the defect engineering layer may involve depositing a layer of the first metal with a suitable thickness, such as a thickness between about 0.2nm and about 5nm, a thickness between about 0.5nm and about 2nm, etc.

[00116] At block 1320, a second layer including a second metallic material may be fabricated on the first layer of the first metallic materials. The second metallic material may also contribute to electronic defect engineering and generating in electronic defects in the switching

29 oxides. The second metallic material may include a second metallic element that is different from the first metallic element. For example, the second metallic element may be Ta. In some embodiments, fabricating the second layer including the second metallic material may involve depositing a second metal (e.g., Ta metal) utilizing PVD, CVD, sputtering, ALD, and/or any other suitable deposition technique. Fabricating the second layer of the second metal may involve depositing a layer of the second metal with a suitable thickness, such as a layer of the second metal that is thicker than that of the first layer of the first metal. In some embodiments, a layer of the second metal having a thickness between lOnm and lOOnm may be deposited. In some embodiments, the second layer of the second metal may be deposited directly on the first layer of the first metal. In such embodiments, a surface of the first layer of the first metal may directly contact one or more portions of a surface of the second layer of the second metal.

[00117] In some embodiments, fabricating the second layer including the second metallic material may involve fabricating a layer including one or more alloys. As described above, each of the alloys may contain a first metallic element and one or more second metallic elements.

Each of the second metallic elements may have different reactivity to the transition metal oxide in the switching oxide layer than the first metallic element. In some embodiments, the first metallic element may be Ta. The second metallic elements may be one or more of W, Hf, Mo, Nb, Zr, etc. Fabricating the second layer of the Ta alloy may involve depositing the Ta alloy utilizing PVD, CVD, sputtering, ALD, and/or any other suitable deposition technique. Fabricating the second layer of the Ta alloy may involve depositing a layer of the Ta alloy with a suitable thickness, such as a layer of the Ta alloy that is thicker than that of the first layer of the first metal. In some embodiments, a layer of one or more Ta alloys having a thickness between about 5nm and about lOOnm may be deposited.

[00118] For simplicity of explanation, the methods of this disclosure are depicted and described as a series of acts. However, acts in accordance with this disclosure can occur in various orders and/or concurrently, and with other acts not presented and described herein. Furthermore, not all illustrated acts may be required to implement the methods in accordance with the disclosed subject matter. In addition, those skilled in the art will understand and appreciate that the methods could alternatively be represented as a series of interrelated states via a state diagram or events.

[00119] FIGS. 15A, 15B, 15C, and 15D illustrate multilevel resistance and linearity data of example RRAM devices in accordance with some embodiments of the present disclosure.

FIG. 15A illustrates current-voltage characteristics 1500A of an RRAM device in the lkD-lOkD

30 range. As shown, according to conductance quantum Go (12.9kQ), the device resistance R is not greater than Go and the device is operated in a filamentary mode (e.g., as shown in FIG 3B).

The conduction of the RRAM device may present metallic behavior.

[00120] FIG. 15B illustrates current-voltage characteristics 1500B of an example RRAM device with high resistance (e.g., 10kQ-100kQ) in accordance with one implementation of the present disclosure. The RRAM device is operated in a transition from a filamentary mode (e.g., with a resistance of lOkG that is not greater than Go) to a non-filamentary mode (e.g., with a resistance of lOOkG that is greater than Go), or a transition from the RRAM device shown in FIG 3B to the RRAM device shown in 3C.

[00121] FIG. 15C illustrates current-voltage characteristics 1500C of an example RRAM device with high resistance (e.g., lOOkG- 1 MW) in accordance with another implementation of the present disclosure. The RRAM device is completely operated in a non-filamentary mode.

The conduction of the RRAM device may present semiconductor behavior. However, through electronic defect engineering including non-continuous interface layer 532, Ta-alloys as the second electrode 540 and/or bi-layer metal as the second electrode 710/720, multilevel high resistances and linearities are demonstrated.

[00122] FIG. 15D illustrates current-voltage characteristics 1500d of an example RRAM device with very high resistance (e.g., 1MW-10MW) in accordance with an implementation of the present disclosure. Implementing such a high resistance in the high resistance range of the non- filamentary mode may require a balanced electronic defect engineering for the RRAM device.

As shown, the RRAM devices with engineered electronic defects present excellent linearity and analog behaviors in this high resistance range.

[00123] The terms “approximately,” “about,” and “substantially” as used herein may mean within a range of normal tolerance in the art, such as within 2 standard deviations of the mean, within ±20% of a target dimension in some embodiments, within ±10% of a target dimension in some embodiments, within ±5% of a target dimension in some embodiments, within ±2% of a target dimension in some embodiments, within ±1% of a target dimension in some embodiments, and yet within ±0.1% of a target dimension in some embodiments. The terms “approximately” and “about” may include the target dimension. Unless specifically stated or obvious from context, all numerical values described herein are modified by the term “about.”

[00124] As used herein, a range includes all the values within the range. For example, a range of 1 to 10 may include any number, combination of numbers, sub-range from the numbers of 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10 and fractions thereof.

31 [00125] In the foregoing description, numerous details are set forth. It will be apparent, however, that the disclosure may be practiced without these specific details. In some instances, well-known structures and devices are shown in block diagram form, rather than in detail, in order to avoid obscuring the disclosure.

[00126] The terms “first,” “second,” “third,” “fourth,” etc. as used herein are meant as labels to distinguish among different elements and may not necessarily have an ordinal meaning according to their numerical designation.

[00127] The words "example" or "exemplary" are used herein to mean serving as an example, instance, or illustration. Any aspect or design described herein as "example" or "exemplary" is not necessarily to be construed as preferred or advantageous over other aspects or designs. Rather, use of the words "example" or "exemplary" is intended to present concepts in a concrete fashion. As used in this application, the term "or" is intended to mean an inclusive "or" rather than an exclusive "or". That is, unless specified otherwise, or clear from context, "X includes A or B" is intended to mean any of the natural inclusive permutations. That is, if X includes A; X includes B; or X includes both A and B, then "X includes A or B" is satisfied under any of the foregoing instances. In addition, the articles "a" and "an" as used in this application and the appended claims should generally be construed to mean "one or more" unless specified otherwise or clear from context to be directed to a singular form. Reference throughout this specification to "an implementation" or "one implementation" means that a particular feature, structure, or characteristic described in connection with the implementation is included in at least one implementation. Thus, the appearances of the phrase "an implementation" or "one implementation" in various places throughout this specification are not necessarily all referring to the same implementation.

[00128] As used herein, when an element or layer is referred to as being “on” another element or layer, the element or layer may be directly on the other element or layer, or intervening elements or layers may be present. In contrast, when an element or layer is referred to as being “directly on” another element or layer, there are no intervening elements or layers present.

[00129] Whereas many alterations and modifications of the disclosure will no doubt become apparent to a person of ordinary skill in the art after having read the foregoing description, it is to be understood that any particular embodiment shown and described by way of illustration is in no way intended to be considered limiting. Therefore, references to details of

32 various embodiments are not intended to limit the scope of the claims, which in themselves recite only those features regarded as the disclosure.

33