阅读说明：本技术 磁阻式随机存取存储器（mram）及其制造方法 (Magnetoresistive Random Access Memory (MRAM) and method of manufacturing the same ) 是由 庄学理 王宏烵 蔡俊佑 黄胜煌 于 2019-02-18 设计创作，主要内容包括：一些实施例涉及包括磁阻式随机存取存储器(MRAM)单元的集成电路及其制造方法。集成电路包括下金属层和设置在下金属层上方的上金属层。底电极设置在下金属层上方并且与下金属层电接触。磁隧道结(MTJ)设置在底电极的上表面上方。顶电极设置在MTJ的上表面上方,并且与上金属层接触。侧壁间隔件围绕顶电极的外周。蚀刻停止层设置在间隔件顶面的外周的顶部上并且围绕上金属层的底面的外周。蚀刻停止层悬于间隔件顶面的外周之上。(Some embodiments relate to integrated circuits including Magnetoresistive Random Access Memory (MRAM) cells and methods of fabricating the same. The integrated circuit includes a lower metal layer and an upper metal layer disposed above the lower metal layer. A bottom electrode is disposed over and in electrical contact with the lower metal layer. A Magnetic Tunnel Junction (MTJ) is disposed over an upper surface of the bottom electrode. The top electrode is disposed over an upper surface of the MTJ and is in contact with the upper metal layer. The sidewall spacer surrounds the periphery of the top electrode. An etch stop layer is disposed on top of the periphery of the top surface of the spacer and around the periphery of the bottom surface of the upper metal layer. The etch stop layer overhangs the periphery of the top surface of the spacer.)

1. An integrated circuit, comprising:

A semiconductor substrate;

An interconnect structure disposed over the semiconductor substrate and including a plurality of dielectric layers and a plurality of metal layers stacked on top of each other in an alternating manner, wherein the plurality of metal layers includes a lower metal layer and an upper metal layer disposed over the lower metal layer;

A bottom electrode disposed above and in electrical contact with the lower metal layer;

a Magnetic Tunnel Junction (MTJ) disposed above an upper surface of the bottom electrode;

A top electrode disposed above an upper surface of the magnetic tunnel junction, wherein the top electrode has an electrode top surface in direct electrical contact with a bottom surface of the upper metal layer;

A sidewall spacer surrounding a periphery of the top electrode, wherein the sidewall spacer has a spacer top surface;

an etch stop layer disposed on top of a periphery of the spacer top surface and surrounding a periphery of the bottom surface of the upper metal layer; and

Wherein the etch stop layer includes a lateral extension overhanging an outer periphery of the spacer top surface.

2. the integrated circuit of claim 1, wherein a bottom surface of the upper metal layer is in contact with the spacer top surface.

3. The integrated circuit of claim 1, wherein the bottom surface has a width less than a width of the top surface of the spacer.

4. The integrated circuit of claim 1, wherein the magnetic tunnel junction has sidewalls that can be sloped at an angle other than 90 degrees as measured from a normal through the upper surface of the bottom electrode.

5. a Magnetoresistive Random Access Memory (MRAM) cell disposed on a semiconductor substrate, the magnetoresistive random access memory cell comprising:

a bottom electrode disposed over the semiconductor substrate;

A Magnetic Tunnel Junction (MTJ) disposed above the bottom electrode;

A top electrode disposed above an upper surface of the magnetic tunnel junction, wherein the top electrode has an electrode top surface;

A sidewall spacer surrounding a periphery of the top electrode, wherein the sidewall spacer has a spacer top surface;

An etch stop layer disposed on top of a periphery of the top surface of the spacer, wherein the etch stop layer overhangs the periphery of the top surface of the spacer; and

A metal line disposed over the top electrode and having a bottom surface in direct physical and electrical contact with the top surface of the electrode.

6. A magnetoresistive random access memory cell according to claim 5 wherein a bottom surface of the metal line is in contact with the spacer top surface.

7. A method for fabricating a Magnetoresistive Random Access Memory (MRAM) cell, the method comprising:

Forming an etch stop layer over an upper surface of the dielectric layer, wherein the etch stop layer has an opening that exposes at least a portion of an upper surface of an underlying metal line;

Forming a bottom electrode layer over the etch stop layer, the bottom electrode layer extending down through the opening to physically and electrically connect to the underlying metal line;

Forming a Magnetic Tunnel Junction (MTJ) layer over the bottom electrode layer;

forming a top electrode over the magnetic tunnel junction layer;

Forming a spacer layer surrounding at least the magnetic tunnel junction layer and the top electrode;

Etching the spacer layer to expose an electrode top surface of the top electrode and a spacer top surface of the spacer layer,

Forming an upper etch stop layer over the electrode top surface and the spacer top surface, wherein the upper etch stop layer overhangs an outer perimeter of the spacer top surface; and

And forming an upper metal layer in contact with the top surface of the electrode.

8. The method of claim 7, wherein portions of the etch stop layer extending beyond the periphery of the spacer top surface slope slightly downward toward the bottom electrode layer.

9. the method of claim 7, wherein the upper etch stop layer has a width greater than a width of a bottom surface of the upper metal layer.

10. The method of claim 9, wherein the upper etch stop layer comprises silicon nitride (Si3N 4).

Technical Field

Embodiments of the invention relate generally to the field of semiconductors, and more particularly, to Magnetoresistive Random Access Memory (MRAM) and methods of fabricating the same.

Background

Many modern electronic devices contain electronic memory. The electronic memory may be volatile memory or non-volatile memory. Non-volatile memories are capable of retaining their stored data without power, while volatile memories lose their stored data when power is removed. Magnetoresistive Random Access Memory (MRAM) is a promising candidate for next generation electronic memory, due to its advantages over current electronic memories. MRAM is generally faster and has better endurance than current non-volatile memories, such as flash random access memory. MRAM typically has similar performance and density, but lower power consumption than current volatile memories, such as Dynamic Random Access Memory (DRAM) and Static Random Access Memory (SRAM).

disclosure of Invention

according to an aspect of the invention, there is provided an integrated circuit comprising: a semiconductor substrate; an interconnect structure disposed over the semiconductor substrate and including a plurality of dielectric layers and a plurality of metal layers stacked on top of each other in an alternating manner, wherein the plurality of metal layers includes a lower metal layer and an upper metal layer disposed over the lower metal layer; a bottom electrode disposed above and in electrical contact with the lower metal layer; a Magnetic Tunnel Junction (MTJ) disposed above an upper surface of the bottom electrode; a top electrode disposed above an upper surface of the magnetic tunnel junction, wherein the top electrode has an electrode top surface in direct electrical contact with a bottom surface of the upper metal layer; a sidewall spacer surrounding a periphery of the top electrode, wherein the sidewall spacer has a spacer top surface; an etch stop layer disposed on top of a periphery of the spacer top surface and surrounding a periphery of the bottom surface of the upper metal layer; and wherein the etch stop layer comprises a lateral extension overhanging an outer periphery of the spacer top surface.

According to another aspect of the present invention, there is provided a Magnetoresistive Random Access Memory (MRAM) cell disposed on a semiconductor substrate, the magnetoresistive random access memory cell comprising: a bottom electrode disposed over the semiconductor substrate; a Magnetic Tunnel Junction (MTJ) disposed above the bottom electrode; a top electrode disposed above an upper surface of the magnetic tunnel junction, wherein the top electrode has an electrode top surface; a sidewall spacer surrounding a periphery of the top electrode, wherein the sidewall spacer has a spacer top surface; an etch stop layer disposed on top of a periphery of the top surface of the spacer, wherein the etch stop layer overhangs the periphery of the top surface of the spacer; and a metal line disposed over the top electrode and having a bottom surface in direct physical and electrical contact with the top surface of the electrode.

According to yet another aspect of the present invention, there is provided a method for fabricating a Magnetoresistive Random Access Memory (MRAM) cell, the method comprising: forming an etch stop layer over an upper surface of the dielectric layer, wherein the etch stop layer has an opening that exposes at least a portion of an upper surface of an underlying metal line; forming a bottom electrode layer over the etch stop layer, the bottom electrode layer extending down through the opening to physically and electrically connect to the underlying metal line; forming a Magnetic Tunnel Junction (MTJ) layer over the bottom electrode layer; forming a top electrode over the magnetic tunnel junction layer; forming a spacer layer surrounding at least the magnetic tunnel junction layer and the top electrode; etching the spacer layer to expose an electrode top surface of the top electrode and a spacer top surface of the spacer layer, an upper etch stop layer being formed over the electrode top surface and the spacer top surface, wherein the upper etch stop layer overhangs an outer periphery of the spacer top surface; and forming an upper metal layer in contact with the top surface of the electrode.

Drawings

Various aspects of the invention are best understood from the following detailed description when read with the accompanying drawing figures. It should be noted that, in accordance with standard practice in the industry, various components are not drawn to scale. In fact, the dimensions of the various elements may be arbitrarily increased or decreased for clarity of discussion.

Fig. 1A illustrates a cross-sectional view of a portion of an electronic memory that includes some embodiments of an MRAM cell that includes a Magnetic Tunnel Junction (MTJ).

Fig. 1B shows a cross-sectional view of an MRAM cell showing the geometry of a stop layer deposited during fabrication of the MRAM cell.

Fig. 1C shows a cross-sectional view of an MRAM cell that exhibits undesirable metal overflow.

Fig. 2 illustrates a cross-sectional view of some embodiments of an integrated circuit including MRAM cells.

Fig. 3 illustrates a top view of some embodiments of the integrated circuit of fig. 2 including MRAM cells.

Fig. 4 shows an enlarged cross-sectional view of an MRAM cell of the integrated circuit of fig. 2.

fig. 5-11 show a series of cross-sectional views of a series of incremental manufacturing steps.

Fig. 12 illustrates a method in flow chart form showing some embodiments of the inventive concept.

Detailed Description

The following disclosure provides many different embodiments, or examples, for implementing different features of the invention. Specific examples of components and arrangements are described below to simplify the present disclosure. These are, of course, merely examples and are not intended to limit the invention. For example, in the following description, forming a first feature over or on a second feature may include embodiments in which the first and second features are formed in direct contact, and may also include embodiments in which additional features may be formed between the first and second features, such that the first and second features may not be in direct contact. Further, the present invention may repeat reference numerals and/or characters in the various embodiments. This repetition is for the purpose of simplicity and clarity and does not in itself dictate a relationship between the various embodiments and/or configurations discussed.

Also, spatially relative terms, such as "below …," "below …," "lower," "above …," "upper," and the like, may be used herein for ease of description to describe one element or component's relationship to another element (or other) component as illustrated. Spatially relative terms are intended to encompass different orientations of the device in use or operation in addition to the orientation depicted in the figures. The device may be otherwise oriented (rotated 90 degrees or at other orientations) and the spatially relative descriptors used herein interpreted accordingly.

a Magnetoresistive Random Access Memory (MRAM) cell includes upper and lower electrodes and a Magnetic Tunnel Junction (MTJ) disposed between the upper and lower electrodes. In a conventional MRAM cell, the upper electrode is connected to an overlying metal layer (e.g., metal 1, metal 2, metal 3, etc.) through a contact or via. Although such connecting contacts or vias are widely employed, the overall height of the MRAM cell plus the contact or via located above it is relatively large relative to the typical vertical spacing between adjacent metal layers (e.g., between metal 2 and metal 3 layers). To make this height more consistent with the vertical spacing between adjacent metal layers, the present invention provides a technique of connecting the top electrode directly to the overlying metal line without a via or contact between the top electrode and the metal line, while avoiding MRAM shorts that may result from the metal line extending beyond the top surface of the MRAM cell and the bottom electrode of the MRAM cell.

Referring to fig. 1A, fig. 1A is a cross-sectional view of a portion of a memory device 100 including a memory array region and a peripheral region. According to some embodiments, the storage region comprises a metal layer to metal layer connection arrangement 103 for the MRAM cell 101. Two MRAM cells 100 (cell 1 and cell 2) are shown, but for convenience the same reference numerals are used to describe MRAM cell 101. The MRAM cell 101 includes a bottom electrode 102 and a top electrode 104 separated from each other by a Magnetic Tunnel Junction (MTJ) 106. In some embodiments, the bottom electrode 102 employs a multi-layer structure (e.g., three layers) including a barrier layer of tantalum nitride or tantalum and two additional layers of tantalum nitride or titanium nitride. A portion of the top electrode 104, MTJ 106, and bottom electrode 102 are surrounded by sidewall spacers 126. The bottom electrode 102 and the top electrode 104 are disposed between the lower metal layer 114 and the upper metal layer 116. The sidewall spacers 126 are surrounded by a protective layer 125, the protective layer 125 may be made of, for example, silicon oxynitride (e.g., SiON), and a dielectric material such as an interlayer dielectric (ILD) or inter-metal dielectric (IMD) layer 128 surrounds the protective layer 125. A dielectric liner 138, such as a silicon dioxide liner or a silicon nitride liner, may be conformally disposed over the dielectric protection layer 140. The dielectric protective layer 140 electrically isolates the bottom electrode 102 from other active circuitry and provides mechanical and chemical protection for the bottom electrode. In some embodiments, the dielectric protection layer is made of silicon dioxide (SiO2) or silicon nitride (Si3N 4).

The MTJ 106 includes a lower ferromagnetic electrode 108 and an upper ferromagnetic electrode 110 separated from each other by a tunneling barrier layer 112. In some embodiments, lower ferromagnetic electrode 108 may have a fixed or "pinned" magnetic orientation, while upper ferromagnetic electrode 110 has a variable or "free" magnetic orientation and may be switched between two or more different magnetic polarities, each magnetic polarity representing a different data state, such as a different binary state. However, in other embodiments, the MTJ 106 may be "flipped" vertically such that the lower ferromagnetic electrode 108 has a "free" magnetic orientation and the upper ferromagnetic electrode 110 has a "fixed" magnetic orientation.

In some embodiments, the sidewall spacers 126 include a spacer top surface 126a that is at substantially the same height as the electrode top surface 104a of the top electrode 104. Portions of etch stop layer 142a remain disposed atop the spacer top surface and around the outer perimeter of upper metal layer 116. The etch stop layer 142a has a width d1, which is a factor of the width d2 defining the bottom surface of the upper metal layer 116 d 1. The width d1 of etch stop layer 142a is controlled, in part, by the width of spacer top surface 126a, which spacer top surface 126a supports etch stop layer 142a as etch stop layer 142a is deposited. It can be seen that the lower portion of etch stop layer 142b extends outwardly from the bottom of sidewall spacer 126.

Fig. 1B schematically illustrates how the width of spacer top surface 126a controls the width of etch stop layer 142a in MRAM cell 150 in some embodiments. In some embodiments, the etch stop layers 142a ', 142 b' may be made of silicon carbide (SiC). The upper portion 142 a' of the etch stop layer may include a central region directly above (and in some cases directly contacting) the upper electrode 104, and a peripheral region that tapers or slopes downward above the spacer 126. It can be seen that etch stop layer 142 a' extends slightly beyond the edges of sidewall spacers 126. The etch stop layer 142 a' has a "beret" shape because the etch stop layer includes lateral extensions that overhang the spacers 126 at significant angles. The portion of the etch stop layer 142 a' that extends beyond the outer perimeter of the top surface of the spacer slopes slightly downward toward the bottom metal layer. For the purposes of this description, the term "overhanging etch stop layer" will be used as a shorthand for describing a beret-shaped etch stop layer configured as shown in fig. 1B. The overhanging etch stop layer 142 a' may prevent unintentional etching of the protective layer 125 in areas extending beyond the outer perimeter of the sidewall spacer 126. When etch stop layer 142a 'is etched to form an opening for the upper metal layer, the opening will not extend beyond etch stop layer 142 a', thereby containing the upper metal layer within the opening and confining the upper metal layer to the area above the MRAM cell, as can be seen in fig. 1A.

In some MRAM fabrication processes, a titanium/titanium nitride layer is deposited on top of the top electrode 104 to prevent oxidation during fabrication. The titanium/titanium nitride layer is removed by a subsequent photo/etch step. An advantage of depositing the stop layer 142a 'on top of the top electrode 104 is that the complete coverage of the stop layer 142 a' over the top electrode 104 may be substantially protected from oxidation and thus a titanium/titanium nitride layer may not be required. Thus, using the etch stop layer 142 a' instead of the titanium/titanium nitride layer to prevent oxidation may save process steps and cost.

Fig. 1C illustrates an exemplary MRAM cell 160 that exhibits one potential difficulty presented by a top electrode 104 'in direct contact with an overlying metal layer 116' without a sufficiently wide sidewall spacer or stop layer. Sidewall spacers 126' are narrower than sidewall spacers 126 of fig. 1B. This means that the etch stop layer 142a "lacks lateral coverage (e.g., width) and may not provide sufficient protection to prevent inadvertent etching of the protective layer 125. During the etch that forms the opening for the overlying metal layer 116 ", if the etch extends slightly beyond the sidewall spacer 126', an unintended cavity may be formed. If the cavity portion is filled with an upper metal layer, a "tooth" portion 116X is formed and a defect (represented by the dashed arrow labeled X) may be created, which represents the possibility of a short circuit between the tooth portion 116X and the bottom electrode 102' of the MRAM cell 160.

Returning to fig. 1A, MRAM cell 100 includes a wide sidewall spacer 126 and an etch stop layer 142a having a sufficient width such that the connection between metal layer 116 and MRAM cell 100 does not extend beyond a top surface 126a of sidewall spacer 126. This means that the risk of a short circuit between the bottom electrode 102 and the overlying metal layer 116 is reduced. It will be appreciated that the components of fig. 1A may provide reduced spacing between the lower metal layer 114 and the upper metal layer 116 due to direct contact between the top electrode 104 and the upper metal layer 116 without intervening vias, and may also be suitable for simplifying manufacturing techniques.

notably, the top electrode 104 itself is in direct electrical contact with the overlying metal layer 116, rather than contacts or vias connecting the top electrode 104 to the overlying metal layer 116. In some embodiments, the upper metal layer 116 is a metal line or a metal layer jumper. In some embodiments, the bottom surface of the overlying metal layer 116 contacts the top surface 104a of the top electrode 104 and also portions of the top surface 126a of the sidewall spacer 126 at a planar interface. Because there is no via or contact between the top electrode 104 and the overlying metal layer 116, the overall height of the MRAM cell 100 is more readily compatible with back end of line (BEOL) process flows.

Fig. 2 illustrates a cross-sectional view of some embodiments of an integrated circuit 200, the integrated circuit 200 including MRAM cells 202a, 202b disposed in an interconnect structure 204 of the integrated circuit 200. Integrated circuit 200 includes a substrate 206. For example, the substrate 206 may be, for example, a bulk substrate (e.g., a bulk silicon substrate) or a silicon-on-insulator (SOI) substrate. The illustrated embodiment shows one or more Shallow Trench Isolation (STI) regions 208, which regions 208 may comprise dielectric-filled trenches within the substrate 206.

two word line transistors 210, 212 are disposed between the STI regions 208. The word line transistor 210 includes a word line gate electrode 214; a word line gate dielectric 218; word line sidewall spacers 222; and source/drain regions 224, and the word line transistor 212 includes a word line gate electrode 216; a word line gate dielectric 220; word line sidewall spacers 222; and source/drain regions 224. Source/drain regions 224 are disposed within the substrate 206 between the word line gate electrodes 214, 216 and the STI regions 208 and are doped to have a first conductivity type opposite to a second conductivity type of the channel regions underlying the gate dielectrics 218, 220, respectively. The word line gate electrodes 214, 216 may be, for example, doped polysilicon or a metal, such as aluminum, copper, or a combination thereof. The word line gate dielectrics 218, 220 may be, for example, an oxide such as silicon dioxide or a high-k dielectric material. The word line sidewall spacers 222 may be made of, for example, silicon nitride (e.g., Si3N 4).

the interconnect structure 204 is disposed over the substrate 206 and connects devices (e.g., transistors 210, 212) to each other. Interconnect structure 204 includes a plurality of IMD layers 226, 228, 230 and a plurality of metallization layers 232, 234, 236 stacked on top of each other in an alternating manner. The IMD layers 226, 228, 230 may be made of, for example, low-k dielectrics such as undoped silicate glass or oxides such as silicon dioxide or very low-k dielectric layers. The metallization layers 232, 234, 236 include metal lines 238, 240, 241, 242 formed within the trenches and may be made of a metal such as copper or aluminum. Contacts 244 extend from the bottom metallization layer 232 to the source/drain regions 224 and/or the gate electrodes 214, 216; and vias 246 extend between the metallization layers 232, 234, 236. The contacts 244 and vias 246 extend through the dielectric protective layers 250, 252 (which may be made of a dielectric material and may serve as an etch stop layer during fabrication). For example, the dielectric protection layers 250, 252 may be made of an extremely low-k dielectric material such as SiC. For example, the contacts 244 and vias 246 may be made of a metal such as copper or tungsten.

MRAM cells 202a, 202b configured to store respective data states are disposed within interconnect structure 204 between adjacent metal layers. MRAM cell 202a includes a bottom electrode 254 and a top electrode 256 that are made of a conductive material. Between the top electrode 256 and the bottom electrode 254 of the MRAM cell 202a, the MRAM cell 202a includes an MTJ 258. MRAM cell 202a also includes sidewall spacer 260. The metal line 242 has a lowermost surface that is coplanar and in direct electrical contact (ohmic connection) with portions of the top surface of the top electrode 256 and the top surface of the sidewall spacer 260.

Fig. 3 illustrates a top view of some embodiments of the integrated circuit 200 of fig. 2 as represented by cut lines in fig. 2-3. It can be seen that the MRAM cells 202a, 202b may have a square, rectangular, or circular shape when viewed from top to bottom in some embodiments. However, in other embodiments, the corners of the illustrated squares may be rounded, such that the MRAM cells 202a, 202b have square or rectangular shapes with rounded corners, or have circular or elliptical shapes, for example, due to the practicality of many etching processes. The MRAM cells 202a, 202b are disposed on metal lines 240, 241, respectively, and have top electrodes 256 that are electrically connected directly to metal line 242 without a via or contact therebetween, respectively.

Referring now to fig. 4, an enlarged cross-sectional view of the MRAM cell 202a of fig. 2 is provided. As shown, MRAM cell 202a includes a bottom electrode 254 and a top electrode 256, wherein an MTJ 258 is disposed between bottom electrode 254 and top electrode 256. The bottom electrode 254 extends down through an opening in the dielectric protection layer 252 to make electrical contact with the underlying metal line 240.

In the illustrated embodiment, MTJ 258 includes a lower ferromagnetic electrode 266 (which may have a fixed magnetic orientation) and an upper ferromagnetic electrode 268 (which may have a free magnetic orientation). A tunneling barrier layer 270 is disposed between the lower ferromagnetic electrode 266 and the upper ferromagnetic electrode 268; and a capping layer 272 is disposed over the upper ferromagnetic electrode 268. The lower ferromagnetic electrode 266 may be a Synthetic Antiferromagnetic (SAF) structure including a top fixed ferromagnetic layer 274, a bottom fixed ferromagnetic layer 276, and a metal layer 278 sandwiched between the top fixed ferromagnetic layer 274 and the bottom fixed ferromagnetic layer 276.

In some embodiments, upper ferromagnetic electrode 268 includes Fe, Co, Ni, FeCo, CoNi, CoFeB, FeB, FePt, FePd, and the like. In some embodiments, capping layer 272 comprises WO2, NiO, MgO, Al2O3, Ta2O5, MoO2, TiO2, GdO, Al, Mg, Ta, Ru, and the like. In some embodiments, the tunnel barrier layer 270 provides electrical isolation between the upper ferromagnetic electrode 268 and the lower ferromagnetic electrode 266, while still allowing electrons to pass through the tunnel barrier layer 270 under appropriate conditions. The tunnel barrier layer 270 may include, for example, magnesium oxide (MgO), aluminum oxide (e.g., Al2O3), NiO, GdO, Ta2O5, MoO2, TiO2, WO2, and the like.

In operation, the variable magnetic polarity of the upper (e.g., free) ferromagnetic electrode 268 is typically read by measuring the resistance of the MTJ 258. The resistance of MTJ 258 varies with variable magnetic polarity due to the magnetic tunneling effect. Further, in operation, the variable magnetic polarity is typically changed or switched using the Spin Transfer Torque (STT) effect. According to the STT effect, current flows through MTJ 258 to induce a flow of electrons from lower (e.g., fixed) ferromagnetic electrode 266 to upper (e.g., free) ferromagnetic electrode 268. As the electrons pass through the lower ferromagnetic electrode 266, the spins of the electrons are polarized. When the spin-polarized electrons reach the upper ferromagnetic electrode 268, the spin-polarized electrons apply a torque to the variable magnetic polarity and switch the state of the free ferromagnetic electrode (e.g., upper electrode 268). Alternative methods for reading or changing the variable magnetic polarity are also acceptable. For example, in some alternative approaches, the magnetization polarity of fixed ferromagnetic electrode 266 and/or free ferromagnetic electrode 268 is perpendicular to the interface between tunneling barrier layer 270 and fixed ferromagnetic electrode 266 and/or free ferromagnetic electrode 268, making MTJ 258 a perpendicular MTJ.

In the illustrated embodiment, because the top electrode 256 itself (and portions of the sidewall spacers 260) are in direct contact with the overlying metal line 242, the overall height of the MRAM cells 202a, 202b may be small relative to previous approaches. This small height makes the MRAM cells 202a, 202b more compatible with BEOL process flows. Thus, the formation of MRAM cells 202a, 202b provides better MRAM operation while reducing manufacturing costs. Furthermore, because the bottom surface of the metal line is not as wide as the top surface of the spacer 260, the likelihood of the metal line shorting to the bottom electrode 254 is reduced.

referring to fig. 5-11, cross-sectional views of some embodiments of semiconductor structures having MRAM cells at various stages of fabrication are provided. While fig. 5-11 are described as a series of steps, it should be understood that these steps are not limiting, the order of the steps may be varied in other embodiments, and the disclosed methods are applicable to other configurations as well. In other embodiments, some steps shown and/or described may be omitted, in whole or in part.

Figure 5 illustrates a cross-sectional view showing some embodiments of a portion of an interconnect structure 204 disposed over a substrate (not shown in figure 5, but previously shown in figure 2). Interconnect structure 204 includes IMD layer 228 and metal line 240 extending horizontally through IMD layer 228. IMD layer 228 may be an oxide such as silicon dioxide, a low-k dielectric material, or an ultra-low-k dielectric material. The metal line 240 may be made of metal such as aluminum, copper, or a combination thereof. In some embodiments, the substrate may be a bulk silicon substrate or a semiconductor-on-insulator (SOI) substrate (e.g., silicon-on-insulator). For example, the substrate may also be a binary semiconductor substrate (e.g., GaAs), a ternary semiconductor substrate (e.g., AlGaAs), or a higher order semiconductor substrate. In many instances, the substrate represents a semiconductor wafer, and may have a thickness of 1 inch (25 mm); 2 inches (51 mm); 3 inches (76 mm); 4 inches (100 mm); 5 inches (130mm) or 125mm (4.9 inches); 150mm (5.9 inches, commonly referred to as "6 inches"); 200mm (7.9 inches, commonly referred to as "8 inches"); 300mm (11.8 inches, commonly referred to as "12 inches"); 450mm (17.7 inches, commonly referred to as "18 inches") in diameter. After the process is completed, such a wafer may optionally be stacked with other wafers or dies, for example, after formation of MRAM cells, and then singulated into individual dies corresponding to individual ICs.

A first dielectric protection layer 252 is formed over IMD layer 228 and over metal line 240. In some embodiments, the first dielectric protection layer 252 comprises SiC (silicon carbide) having a thickness of about 250 angstroms. A second dielectric protection layer 253 is formed over the first dielectric protection layer 252. In some embodiments, the second dielectric protection layer has a different chemical composition than the first dielectric protection layer 252, and may, for example, comprise SRO (silicon-rich oxide) having a thickness of about 200 angstroms. The bottom electrode layer 254 is formed over the dielectric protection layers 252, 253 and extends down through openings in the dielectric protection layers 252, 253 to make electrical contact with upper portions of the metal lines 240. The bottom electrode layer 254 may be a conductive material such as, for example, titanium nitride, tantalum nitride, titanium, tantalum, or a combination of one or more of the foregoing. Further, in some embodiments, the bottom electrode layer 254 may be, for example, about 10 to 100 nanometers thick.

A Magnetic Tunnel Junction (MTJ) stack 258 is formed over an upper surface of the bottom electrode layer 254 and a top electrode layer 256 is formed over the MTJ stack 258. The top electrode layer 256 may be a conductive material such as, for example, titanium nitride, tantalum nitride, titanium, tantalum, tungsten, or a combination of one or more of the foregoing. Further, the top electrode layer 256 may be, for example, about 10 to 100 nanometers thick. A mask 502 is disposed over the upper surface of the top electrode layer 256. In some embodiments, mask 502 comprises a photoresist mask, but may also be a hard mask such as a nitride marker. In some embodiments, the mask 502 may be a different conductive material than the top electrode layer 256, such as, for example, titanium nitride, tantalum nitride, titanium, tantalum, or a combination of one or more of the above. The sidewalls of the MTJ 258 and/or top electrode 256 may be tilted at an angle other than 90 degrees, measured with respect to a normal through the upper surface of the bottom electrode 254.

sidewall spacer precursor layer 260' is formed over the lateral portion of bottom electrode 254, the sidewalls of MTJ 258, the sidewalls of top electrode 256, and extends over the sidewalls and upper surface of mask 502. In some embodiments, the sidewall spacer precursor layer 260' may be formed by any suitable deposition technique, and is generally conformally formed. In addition, sidewall spacer precursor layer 260' may be formed of, for example, silicon nitride, silicon carbide, Si3N4, SiON, or a combination of one or more of the foregoing. Even more, sidewall spacer precursor layer 260' can be formed to have a thickness of, for example, about 150 to 600 angstroms. A dielectric liner 602, such as a conformal oxide, is then formed over the sidewall spacer precursor layer 260'. The dielectric liner 602 facilitates the spacer etch process implemented in fig. 6.

in fig. 6, a spacer etch process 600 (e.g., an anisotropic etch) has been performed on the sidewall spacer precursor layer 260 ' to etch back the sidewall spacer precursor layer 260 ' to remove the lateral extensions of the sidewall spacer precursor layer 260 ' and the top electrode mask layer 502 to expose the top surface of the top electrode 256 surrounded by the remaining sidewall spacers 260. In some embodiments, after etching, the combined width of the sidewall spacer top surface and the electrode top surface is significantly wider than the expected width of the metal well or trench to be formed in fig. 10 to create a metal line (e.g., greater than 154 nm). Thus, in some embodiments, the width of the sidewall spacer is selected based on the width of the metal line to which the top electrode will be connected. In addition, the spacer etch process cuts the bottom electrode 254 to its final size. In some embodiments, the spacer etch 600 is a unidirectional or vertical etch.

in fig. 7, an etch stop layer is deposited to produce a first portion 142a of the etch stop layer overlying the top surfaces of the electrodes and the top surfaces of the spacers. An additional portion 142b of the etch stop layer, which may be discontinuous relative to the first portion 142a, abuts the periphery of the bottom electrode 254. This discontinuity in the stop layer is due to the stepped coverage characteristics of the stop layer material (e.g., silicon nitride, silicon carbide, Si3N4, SiON, or combinations thereof) which is not typically deposited on the lateral surfaces of the MTJ. In addition, the first portion 142a overhangs the spacer top surface, revealing, in some embodiments, the beret shape shown in fig. 1B to provide additional lateral protection against inadvertent etching beyond the spacer top surface.

In fig. 8, a protective layer 230, such as a silicon oxynitride (SiON) layer or an ultra-low k dielectric layer, is then formed over the etch stop layers 142a, 142b, for example by Chemical Vapor Deposition (CVD), Plasma Vapor Deposition (PVD), spin-on techniques, or thermal oxidation. The protective layer 230 electrically isolates the MRAM cells from other active circuitry and provides mechanical and chemical protection for the MRAM cells. In some embodiments, the top surface of the protection layer 230 is about 1080 angstroms above the surface of the second dielectric protection layer 253. In some embodiments, Chemical Mechanical Planarization (CMP) is then performed on the protective layer 230 to planarize the upper surface of the protective layer 230. After the CMP, a photo mask (not shown) is formed over the protective layer 230, and etching is performed such that the protective layer 230 covers the memory array region and does not cover the peripheral region, as shown in fig. 8.

Next, an IMD or ILD layer 801 made of a dielectric material such as an oxide or ELK dielectric is applied on top of the protection layer 230 in the memory array region and on top of the second dielectric protection layer 253 in the peripheral region. In some embodiments, IMD or ILD layer 801 has a thickness of about 400 angstroms in the memory array region and a thickness of about 1700 angstroms in the peripheral region. An etch stop layer 803 is deposited over the IMD or ILD layer 801. In some embodiments, etch stop layer 803 comprises Tetraethylorthosilicate (TEOS). A nitrogen free anti-reflective layer (NFARL)805 is applied on top of the etch stop layer 803. In some embodiments, NFARL 805 is about 200 angstroms thick. A hard mask layer 807 is applied over NFARL 805. Photolithography techniques are used to pattern the hard mask layer with trench openings that will be used in a dual damascene process to form trenches or openings that will hold the top metal layer. In some embodiments, these openings may be dual damascene openings. In some embodiments, the hard mask layer 807 comprises titanium nitride (TiN) and is about 350 angstroms thick.

In fig. 9, a photoresist layer 909 is applied over the hard mask layer 807. A first trench 915 is etched in the peripheral region.

In FIG. 10, photoresist layer 909 has been removed. One or more etches are then performed to form trench openings 242 'and 243'. In some embodiments, the one or more etches comprise a dual damascene process.

In fig. 11, the trenches and openings are filled with a metal such as aluminum or copper. Thus, in the memory array region, the trenches are filled with metal lines 242, the metal lines 242 having a bottom surface in direct contact with the top surface of the top electrode 256, thereby providing an ohmic connection without a contact or via between the metal lines 242 and the top electrode 256. The bottom surface of the metal line also contacts portion 142a of the stop layer, which reduces the risk of metal overflow beyond the MRAM cell. In some embodiments, the bottom surface of the metal line is in contact with a portion of the stop layer. A CMP operation (shown in dashed lines) is then performed to planarize the upper surface of the metal lines and the upper surface of the dielectric protection layer 801, resulting in the structure of fig. 1A and/or fig. 4.

In another region of the integrated circuit, such as a peripheral region where CMOS logic devices are formed, metal line 242 is connected to underlying metal line 240 through via 243. The interposition of the via 243 between the metal layer 242 and the underlying metal line 240 takes up similar space in the vertical direction of the MRAM cell as compared to the direct connection between the metal line 242 and the top electrode 256. Thus, the direct connection between the metal lines 242 and the top electrodes 256 in the memory array area allows the cell height in the memory array area to be reduced such that the cell height in the memory array area is similar to the cell height in the peripheral area.

Fig. 12 illustrates a method 1200 of forming an MRAM cell having an etch stop layer of sufficient width to prevent unintentional etching beyond the sidewall spacers, in accordance with some embodiments. While this and other methods are illustrated and described herein as a series of steps or events, it will be appreciated that the present invention is not limited by the illustrated ordering or steps. Thus, in some embodiments, the steps may be performed in a different order than shown, and/or may be performed simultaneously. Further, in some embodiments, illustrated steps or events may be subdivided into multiple steps or events, which may be performed at different times or concurrently with other steps or sub-steps. In some embodiments, some illustrated steps or events may be omitted, and other steps or events not illustrated may be included.

For example, in some embodiments, steps 1202 through 1208 may correspond to the structure previously shown in fig. 5. In step 1202, an etch stop layer is formed over an upper surface of a dielectric layer. The etch stop layer has an opening that exposes at least a portion of an upper surface of the underlying metal line. In step 1204, a bottom electrode layer is formed over the etch stop layer. The bottom electrode layer extends down through the opening to make physical and electrical contact with the underlying metal layer. In step 1206, a Magnetic Tunnel Junction (MTJ) layer is formed over the bottom electrode layer. In step 1208, a top electrode layer is formed over the magnetic tunnel junction layer. In step 1210, which may correspond to the example previously shown in fig. 5, a wide spacer layer is formed surrounding at least the MTJ layer and the top electrode. The wide spacer layer is wide enough to support an etch stop layer that prevents etching from inadvertently extending beyond the top surface of the spacer. In step 1212, which may correspond to the example previously shown in fig. 6, the spacer layer is etched to expose the top surfaces of the top electrode and the spacer. In step 1213, which may correspond to the example previously shown in fig. 7, an etch stop layer is formed covering the top surface of the top electrode and the top surface of the spacer. The etch stop layer overhangs an outer periphery of the top surface of the spacer. In step 1214, which may correspond to the example previously shown in fig. 11, an upper metal layer is formed in direct physical and electrical contact with the top surfaces of the electrodes and the top surfaces of the spacers.

Some embodiments relate to integrated circuits including Magnetoresistive Random Access Memory (MRAM) cells. An integrated circuit includes a semiconductor substrate and an interconnect structure disposed over the semiconductor substrate. The interconnect structure includes a plurality of dielectric layers and a plurality of metal layers stacked on top of each other in an alternating manner. The plurality of metal layers includes a lower metal layer and an upper metal layer disposed above the lower metal layer. A bottom electrode is disposed over and in electrical contact with the lower metal layer. A Magnetic Tunnel Junction (MTJ) is disposed over an upper surface of the bottom electrode. The top electrode is disposed over an upper surface of the MTJ and has an electrode top surface in direct electrical contact with the upper metal layer. The sidewall spacer surrounds a periphery of the top electrode and has a spacer top surface. An etch stop layer is disposed on top of the periphery of the top surface of the spacer and around the periphery of the bottom surface of the upper metal layer. The etch stop layer overhangs the periphery of the top surface of the spacer.

In some embodiments, a bottom surface of the upper metal layer is in contact with the spacer top surface.

In some embodiments, the bottom surface has a width less than a width of the top surface of the spacer.

In some embodiments, the magnetic tunnel junction has sidewalls that can be sloped at angles other than 90 degrees measured from a normal through the upper surface of the bottom electrode.

In some embodiments, a portion of the etch stop layer that extends beyond the periphery of the spacer top surface slopes slightly downward toward the lower metal layer.

In some embodiments, the integrated circuit further comprises: an additional portion of the etch stop layer disposed at an outer periphery of the bottom electrode. Other embodiments relate to an MRAM cell disposed on a semiconductor substrate. The MRAM cell includes a bottom electrode disposed above a semiconductor substrate, and a Magnetic Tunnel Junction (MTJ) disposed above the bottom electrode. A top electrode is disposed over an upper surface of the MTJ, where the top electrode has an electrode top surface. A sidewall spacer surrounds a periphery of the top electrode, wherein the spacer has a spacer top surface. A metal line is disposed over the top electrode and has a bottom surface in direct physical and electrical contact with the top surface of the electrode and at least a portion of the top surface of the spacer.

In some embodiments, a bottom surface of the metal line is in contact with a top surface of the spacer.

In some embodiments, the magnetic tunnel junction has sidewalls that can be sloped at angles other than 90 degrees measured from a normal through the upper surface of the bottom electrode.

In some embodiments, a width of a bottom surface of the metal line is less than a width of a top surface of the spacer.

In some embodiments, a portion of the etch stop layer that extends beyond the periphery of the top surface of the spacer slopes slightly downward toward the bottom electrode.

in some embodiments, the magnetoresistive random access memory cell further comprises: an additional portion of the etch stop layer disposed at an outer periphery of the bottom electrode.

other embodiments relate to methods for fabricating MRAM cells. In the method, an etch stop layer is formed over an upper surface of the dielectric layer, wherein the etch stop layer has an opening that exposes at least a portion of an upper surface of an underlying metal line. A bottom electrode layer is formed over the etch stop layer. The bottom electrode layer extends down through the opening to be physically and electrically connected to the underlying metal lines. A Magnetic Tunnel Junction (MTJ) layer is formed over the bottom electrode layer. A top electrode is formed over the magnetic tunnel junction layer. A spacer layer is formed surrounding at least the MTJ layer and the top electrode. The spacer layer is etched to expose a top surface of the top electrode and a top surface of the spacer. An upper metal layer is formed in direct electrical and physical contact with the top surfaces of the electrodes and the top surfaces of the spacers.

in some embodiments, a portion of the etch stop layer that extends beyond the periphery of the spacer top surface slopes slightly downward toward the bottom electrode layer.

In some embodiments, the upper etch stop layer has a width greater than a width of the bottom surface of the upper metal layer.

In some embodiments, the upper etch stop layer comprises silicon nitride (Si3N 4).

In some embodiments, the width of the top surface of the spacer in combination with the width of the top surface of the electrode is greater than about 154 nanometers.

In some embodiments, the spacer layer comprises silicon carbide (SiC).

In some embodiments, the method further comprises: forming a dielectric layer over the spacer top surface and the electrode top surface; and forming a trench opening in the dielectric layer, wherein the trench opening exposes portions of the top electrode surface and the top spacer surface; and filling the trench opening directly adjoining portions of the top surfaces of the electrodes and the top surfaces of the spacers with a conductive material.

In some embodiments, the trench opening exposes a portion of the entire spacer top surface.

It should be understood that in this written description and in the claims that follow, the terms "first," "second," "third," etc. are merely general identifiers used for ease of description to distinguish different elements of a figure or series of figures. The terms do not by themselves connote any temporal order or structural proximity of the elements and are not intended to describe corresponding elements in the different illustrated embodiments and/or embodiments not illustrated. For example, a "first dielectric layer" described in connection with a first figure may not necessarily correspond to a "first dielectric layer" described in connection with another figure, and may not necessarily correspond to a "first dielectric layer" in an embodiment not shown.

The foregoing has outlined features of several embodiments so that those skilled in the art may better understand the aspects of the present invention. Those skilled in the art should appreciate that they may readily use the present disclosure as a basis for designing or modifying other processes and structures for carrying out the same purposes and/or achieving the same advantages of the embodiments introduced herein. Those skilled in the art should also realize that such equivalent constructions do not depart from the spirit and scope of the present disclosure, and that they may make various changes, substitutions, and alterations herein without departing from the spirit and scope of the present disclosure.