TWI859274B - Magnetic tunnel junction stack with data retention - Google Patents

Abstract

Disclosed herein are exemplary magnetic tunnel junction structures for magnetic random access memory applications. A magnetic tunnel junction stack includes a structure blocking layer and a magnetic reference layer. The magnetic reference layer is disposed on the structure blocking layer. The magnetic reference layer is formed from cobalt (Co) or iron (Fe). A tunnel barrier layer is disposed on the magnetic reference layer. The tunnel barrier layer is formed from magnesium (Mg). A magnetic storage layer is disposed in contact with the tunnel barrier layer. The magnetic storage layer includes a first magnetic layer formed from cobalt (Co), iron (Fe), and boron (B). The magnetic storage layer includes a non-magnetic layer formed from one of (Mo), tantalum (Ta), tungsten (W). A second magnetic layer is formed from Co, Fe, and B. The non-magnetic layer is disposed in contact with first magnetic layer and the second magnetic layer.

Description

Magnetic tunneling junction stack with data storage function

Examples of this disclosure generally relate to fabricating magnetic tunneling junction structures for magnetic random access memory (MRAM) applications.

Spin transfer torque magnetic random access memory, or STT-MRAM, employs a magnetic tunneling junction structure in its memory cells in which two ferromagnetic layers are separated from each other by a thin insulating or "dielectric" layer. One of the magnetic layers has a fixed magnetic polarity and the other has a magnetic polarity that can be selectively changed (i.e., switched) between two states. The magnetic layers have a perpendicular magnetic anisotropy in the depth direction of the stack of film layers that includes the magnetic tunneling junction "MTJ". The polarity of the variable polarity layer can be switched between having the same polarity as the fixed polarity layer or a polarity opposite to that of the fixed polarity layer. The resistance across the MTJ is a function of the polarity of the variable polarity layer relative to the fixed polarity layer. For example, when the polarity of the two layers is the same in the depth direction of the MTJ, the resistance across the MTJ is lower, i.e., is assigned a value of 0. When the polarity of the two layers is opposite to each other in the depth direction of the MTJ, the resistance across the MTJ is higher, i.e., is assigned a value of 1. Therefore, the resistance across the memory cell can be used to indicate a value of 1 or 0, and therefore to store binary data values.

The Internet of Things (IoT) includes devices such as cellular phones, tablets, and other portable devices, and wearable devices such as sports trackers, smart watches, and health monitors. These IoT devices capture, process, analyze, and store large amounts of data. Current memory storage approaches include the development of embedded flash (eFlash) memory, which is used for storage in cellular phones, tablets, and other portable devices. This is because eFlash is associated with higher cost, lower reliability, and higher write energy consumption for logic circuits on the same chip. Because eFlash devices and logic devices are formed at the front end of the chip, an additional mask layer must be added to minimize the impact of the eFlash floating gate on the performance of the logic gate. Because eFlash is used in more advanced nodes (i.e., more advanced circuit generations and circuit architectures), the additional cost of the additional mask layer dramatically increases the storage cost of the memory.

Therefore, there is still a need for new memory technologies that can be integrated at the back end of the chip at more advanced nodes. Therefore, there is a need for an improved MTJ stack and a method of manufacturing the same in order to improve data retention while reducing production costs.

The present disclosure generally relates to magnetic tunneling junction structures (i.e., stacks) suitable for magnetic random access memory (MRAM) applications and methods of making the same. In one example, the magnetic tunneling junction stack includes a structural barrier layer and a magnetic reference layer. The magnetic reference layer is disposed on the structural barrier layer. The magnetic reference layer is formed of cobalt (Co) or iron (Fe). A tunneling barrier layer is disposed on the magnetic reference layer. The tunneling barrier layer is formed of magnesium (Mg). A magnetic storage layer is disposed in contact with the tunneling barrier layer. The magnetic storage layer includes a first magnetic layer formed of cobalt (Co), iron (Fe), and boron (B). The magnetic storage layer includes a non-magnetic layer formed of one of molybdenum (Mo), tantalum (Ta), and tungsten (W). The second magnetic layer is formed of Co, Fe, and B. The non-magnetic layer is arranged to contact the first magnetic layer and the second magnetic layer.

In a second example, a magnetic tunneling junction stack includes a first pinning layer. A magnetic reference layer is disposed on the first pinning layer. The magnetic reference layer is formed of Co. A tunneling barrier layer contacts the magnetic reference layer. A magnetic storage layer contacts the tunneling barrier layer. The magnetic storage layer includes a first magnetic layer. The first magnetic layer contacts a non-magnetic layer. The first magnetic layer is formed of Co, iron (Fe) and boron (B). The non-magnetic layer is formed of a transition metal. The second magnetic layer is formed of Co, Fe and B. The non-magnetic layer contacts the first magnetic layer and the second magnetic layer.

In another example, a magnetic tunneling junction stack includes a magnetic reference layer and a tunneling barrier layer. The magnetic reference layer is disposed on the tunneling barrier layer. The magnetic reference layer is formed of cobalt (Co). The magnetic storage layer contacts the tunneling barrier layer. The magnetic storage layer includes a first magnetic layer in contact with a non-magnetic layer. The first magnetic layer is formed of Co, iron (Fe) and boron (B). The non-magnetic layer is formed of at least one of molybdenum (Mo), tantalum (Ta) and tungsten (W). The second magnetic layer is formed of Co, Fe and B. The non-magnetic layer separates the first magnetic layer and the second magnetic layer. The capping layer contacts the magnetic storage layer. The covering layer includes a first covering interlayer, a second covering interlayer, and a third covering interlayer. The second covering interlayer and the first magnetic layer are formed of at least one of the same elements. The third covering interlayer and the non-magnetic layer are formed of at least one of Mo, Ta, or W. The first covering interlayer is formed of an oxide of an alkaline earth metal or an oxide of a transition metal. The second covering interlayer separates the first covering interlayer and the third covering interlayer.

100:MTJ stack

102: Substrate

104: Buffer layer

106: Seed layer

108: First nailing layer

110: Coupling layer/ferromagnetic (SyF) coupling layer

112: Second nailing layer

114: Structural barrier

116: Magnetic reference layer

118: Tunneling barrier

120: Magnetic storage layer

122: Covering layer

124: Hard mask layer

150: x direction/size

160: z direction/size

170:y direction/size

204: Buffer layer

204A: First buffer middle layer

204B: Second buffer middle layer

204C: The third buffer middle layer

204D: Double-decker

208: First nailing layer

208A: First middle layer

208B: Second middle layer

208C: Covering layer

212: Second nailing layer

212A: First middle layer

212B: Second middle layer

212C: Covering layer

220: Magnetic storage layer

220A: First magnetic layer/first magnetic storage layer

220B: Second magnetic layer/second magnetic storage layer

220C: non-magnetic layer

222A: First covering middle layer

222B: Second covering middle layer

222C: The third covering middle layer

222D: Fourth covering middle layer

230: Double layer

232: Double layer

234: Double stacking

236: Double stacking

300: hysteresis curve

304: Magnetic force

400: Magnetic dead layer

404:Magnetic volume

408:Thickness

412:Thickness

998:Curve

999: Magnetic force

In order to be able to understand in detail the manner in which the above-mentioned features of the present disclosure are realized, a more specific description of the present disclosure briefly summarized above may be obtained by referring to examples, some of which are shown in the accompanying drawings. However, it should be noted that the accompanying drawings show only exemplary examples and therefore should not be considered as limiting the scope thereof, as the present invention may allow other equally effective examples.

FIG. 1 is a schematic diagram of an MTJ stack according to an example of the present disclosure.

Figure 2A is an enlarged view of the buffer layer of the MTJ stack in Figure 1.

Figure 2B is an enlarged view of the first pinning layer of the MTJ stack in Figure 1.

Figure 2C is an enlarged view of the second pinning layer of the MTJ stack in Figure 1.

FIG. 2D is an enlarged view of an exemplary magnetic storage layer of the MTJ stack of FIG. 1.

FIG. 2E is a magnified view of an exemplary capping layer of the MTJ stack of FIG. 1.

Figure 3 shows the magnetic coercivity of the magnetic storage layer of the MTJ stack in Figure 1.

FIG. 4 is an enlarged view of an exemplary first magnetic layer of the magnetic storage layer of FIG. 2D.

To facilitate understanding, identical reference numerals have been used, where possible, to denote identical elements that are common to the figures. It is contemplated that elements and features of one example may be beneficially incorporated into other examples without further description.

Examples of the present disclosure generally relate to magnetic tunneling junction stacks for magnetic random access memory (MRAM) applications. More specifically, examples described herein relate to magnetic tunneling junction (MTJ) stacks and spin transfer torque magnetoresistive random access memory (STT MRAM). The MTJ stack is incorporated into a film stack that includes an upper electrode and a lower electrode (not shown) of a logic circuit. The MTJ stack is sandwiched between the upper electrode and the lower electrode and can be used to form a plurality of memory cells used in magnetoresistive random access memory (MRAM). In each MTJ stack of the MRAM, there are two magnetic layers. One magnetic layer has a fixed polarity, and the second magnetic layer has a polarity that can be switched by applying a voltage across the second magnetic layer or applying a current to the second magnetic layer. The resistance across the MRAM varies based on the relative polarity between the first and second magnetic layers. The first layer (i.e., the fixed polarity layer) is referred to herein as a magnetic reference layer. The second magnetic layer (i.e., the switching polarity layer) is referred to herein as a magnetic storage layer. The memory cell formed by the MTJ stack works when a voltage is applied across the memory cell or a current is applied through the memory cell to switch the polarity of the second magnetic layer. In response to applying a voltage of sufficient magnitude, the polarity of the switchable magnetic layer is changed to write or store a value to the memory cell. Additionally, the resistivity of the memory cell can be determined by measuring the current versus voltage across the memory cell at a relatively low voltage below the threshold required to switch the magnetic polarity of the magnetic storage layer. Determining the resistivity of the memory cell with a low voltage allows the memory cell to be read without writing or changing the value of the memory cell.

The MTJ stack disclosed herein increases the time a magnetic storage layer can retain data by increasing the magnetic stiffness in the magnetic storage layer. Magnetic stiffness is a measure of the ability of a ferromagnetic material to withstand an external magnetic field without being demagnetized. MRAM stores information by measuring the polarity difference in each magnetic memory cell. The memory cell is formed of two ferromagnetic plates, each of which can maintain a magnetization, i.e., magnetic polarization, separated by an insulating layer. One of the two plates is set to a specific fixed polarity. Data is stored by changing the magnetization strength of the second ferromagnetic plate to match the magnetization strength of the external magnetic field. The MTJ stack of the present disclosure is capable of storing data in the magnetic storage layer at temperatures exceeding 125 degrees Celsius for 10 years.

The MTJ stack discussed herein is formed using a plurality of deposition chambers to deposit thin film layers on a substrate, and ultimately patterning and etching those deposited film layers. The deposition chambers discussed herein for forming the MTJ stack include physical vapor deposition (PVD) chambers and other processing chambers. Here, the PVD chamber is particularly suitable for forming the plurality of thin film layers of the MTJ stack. However, it should be understood that other processing chambers may be equally suitable for forming the thin film layers of the MTJ stack.

FIG. 1 is a schematic diagram of an MTJ stack 100 according to an example of the present disclosure. An x-direction 150 is orthogonal to the stacking direction of the MTJ stack 100. A y-direction 170 is parallel or aligned with the stacking direction of the MTJ stack 100. The MTJ stack 100 includes a substrate 102, a buffer layer 104, a seed layer 106, a first pinning layer 108, a coupling layer 110, a second pinning layer 112, a structural blocking layer 114, a magnetic reference layer 116, a tunneling barrier layer 118, a magnetic storage layer 120, a capping layer 122, and a hard mask layer 124. The stacking direction is perpendicular to the plane of the substrate 102. In the example shown, the buffer layer 104 is formed on the conductive portion of the substrate 102 (or a conductive film formed on the substrate 102) by sputtering or other suitable techniques. The substrate 102 may include tungsten (W), tantalum nitride (TaN), titanium nitride (Tin), or other metal layers. The buffer layer 104 improves the adhesion of the seed layer 106 to the substrate 102, which helps the formation and performance of subsequently deposited layers of the MTJ stack 100. The buffer layer 104 includes Co _x F _y B _z , Ta, and/or TaN, and is formed in one or more deposition operations in a processing chamber. In one example, the buffer layer 104 is formed in a _PVD chamber using Ar plasma and a sputtering target that is a _{CoxFeyBz alloy} , or by using a single sputtering target of Co, Fe, or B, or by a combination of an alloy sputtering target and a single element sputtering target (e.g., a CoFe target and a B target). In one example, x is an integer such that the molecular weight percentage of _Cox is about 10% to about 40% of the molecular weight of _CoxFeyBz (i.e., _{the compound} ), y is an integer such that the molecular weight percentage of _Fey is 20% to about 60% of the molecular weight of the compound, and z is an integer such that the molecular weight percentage of _Bz is equal to or less than 70% of the molecular weight of the compound. In another example, when the buffer layer 104 includes a Ta layer, the Ta layer may be formed in a PVD chamber using a Ta target and Ar plasma.

In one example, the buffer layer 104 includes TaN, and the buffer layer 104 is formed on the substrate 102 in the processing chamber. The processing chamber may use a Ta target in the presence of Ar plasma and _N2 . _N2 reacts with the Ta material to form a TaN layer. In another example, a TaN sputtering target is used in a PVD chamber with Ar plasma to form the buffer layer 104. In one example, the buffer layer 104 is formed directly on and in contact with the conductive layer on the substrate 102. In other examples, there is a conductive transition layer between the conductive layer on the substrate 102 and the buffer layer 104 that does not affect the performance of the MTJ stack 100. The use of buffer layer 104 is optional in the illustrated example, and in some cases may not be used in other examples discussed herein.

When present, the optional buffer layer 104 has a total thickness of the buffer layer 104 from about 0Å to about 60Å. In one example, the buffer layer 104 is a single layer of Ta, TaN, or Co _x Fe _y B _z formed directly on and in contact with the conductive layer on the substrate 102, and its thickness is up to about 20Å. In another example, the buffer layer 104 is a combination of layers. Each layer of the buffer layer 104 is at least one of Ta, TaN, or Co _x Fe _y B _z . The buffer layer 104 is formed at a thickness between about 1Å and about 60Å. In an example in which TaN is used instead of Ta or Co _x Fe _y B _z for the buffer layer 104, the thickness can be about 20Å. In one example where Co _x F _ey B _z is used alone to form the buffer layer 104, the thickness of the buffer layer 104 may be about 10 Å. In another example of the buffer layer 104, Ta or TaN is used in combination with Co _x F _ey B _z , and the thickness of the buffer layer 104 is about 20 Å. Within the buffer layer 104, the thickness of the Ta and/or TaN layer may be between about 0 Å and about 40 Å. In one example, the TaN layer has a thickness of about 20 Å.

The seed layer 106 is deposited on the buffer layer 104. The seed layer 106 includes at least one of Cr, NiCr, NiCrFe, RuCr, IrCr, CoCr. Here, Co is cobalt, Cr is chromium, Ir is iridium, Fe is iron, and Ru is ruthenium. The seed layer 106 may be formed as one or more layers of Cr, NiCr, NiCrFe, RuCr, IrCr, or CoCr, which may include a combination of those elements or a combination of alloys in a single layer. In one example, the seed layer 106 has a thickness of about 100Å or less. For example, the seed layer 106 has a thickness from about 30Å to about 60Å thick.

The seed layer 106 is formed directly on and in contact with the buffer layer 104. Alternatively, a transition layer may exist between the seed layer 106 and the buffer layer 104. The transition layer is selected to have minimal impact on the performance of the MTJ stack 100.

In one example of the disclosed MTJ stack 100, the seed layer 106 is made of NiCr, the first pinning layer 108 is made of a Co/Pt bilayer stack, and the second pinning layer 112 is made of Co. In a second example, the first pinning layer 108 may be made of a Co/Ni bilayer stack, and the second pinning layer 112 is made of Co. In this example, the MTJ stack 100 exhibits a tunneling magnetoresistance (TMR) of 175% after being formed.

The first pinning layer 108 is formed on the seed layer 106. The first pinning layer 108 can be formed using a physical vapor deposition process. In one example, the first pinning layer 108 is made as a single layer of Co having a thickness of about 1Å to about 18Å. In another example, the first pinning layer 108 is made of one or more bilayers of various materials, where each bilayer may include two or more intermediate layers. The first pinning layer 108 may include one or more bilayers alone or in combination with a Co layer. In an example where one or more double layers are included in the first pinning layer 108, each double layer contains a first intermediate layer of Co and a second intermediate layer including a compound of at least one element other than Co. The first pinning layer 108 may be formed by sputtering a Co target using Ar plasma, and then sputtering a second target of Pt, Ir, Ni, or Pd. In an example where Pt is used with Co to form the double layer, Xe plasma may be used instead of or in addition to Ar plasma. In an example where one or more double layers are used to form the first pinning layer, repeated deposition cycles may be performed in a processing chamber by forming the first intermediate layer of the double layer. In one example, the first intermediate layer includes Co and is formed by shielding a target in a PVD chamber that does not include Co, and shielding the Co target and other targets to expose a second target, the second target including a second element for the second intermediate layer of the double layer. The deposition of the first and second intermediate layers can be repeated in an iterative manner to form one or more double layers of the first pinning layer 108. Optionally, the first intermediate layer can be formed in an ALD, CVD or other deposition chamber (such as a PECVD chamber). In one example, the first pinning layer 108 is formed directly on and in contact with the seed layer 106. In other examples, an optional transition layer is formed between the seed layer 106 and the first pinning layer 108.

A synthetic ferromagnet (SyF) coupling layer 110 is formed on the first pinning layer 108, and a second pinning layer 112 is formed on the SyF coupling layer 110. The SyF coupling layer 110 is formed by a PVD process using a process gas or a target containing Ru, Rh, Cr, or Ir. The thickness of the SyF coupling layer 110 is from about 3Å to about 10Å. In one example, Ru has a thickness from about 4Å to about 5Å. In another example, Ru has a thickness from about 7Å to about 9Å. In another example, the SyF coupling layer 110 includes Ir, and its thickness is from about 4Å to about 6Å. In another example, the second pinning layer 112 is manufactured by sputtering a Co target. In one example, the SyF coupling layer 110 is directly formed on and contacts the first and second pinning layers 108 and 112. In other examples, a transition layer exists between the SyF coupling layer 110 and either or both of the first and second pinning layers 108 and 112 without affecting the performance of the MTJ stack 100.

The second pinning layer 112 may be formed of one or more double layers of Co and another element included in the second pinning layer 112. In one example, the second pinning layer 112 is formed by sputtering a Co target using Ar plasma and then sputtering a second target of Pt, Ir, Ni, or Pd using Ar plasma. Repeated deposition cycles performed in a processing chamber (e.g., using various processing gases or using a Co target and a second target in the presence of Ar plasma) may be used to form one or more double layers of the second pinning layer 112. In one example, the second pinning layer 112 is formed as a single Co layer having a thickness from about 0 Å to about 10 Å. In another example, the thickness of the second pinning layer 112 is about 5Å. An alternative configuration of the second pinning layer 112 is shown in FIG. 2C.

A structural blocking layer 114 is optionally formed on the second pinning layer 112. The structural blocking layer 114 prevents a short circuit from being formed between the MTJ stack 100 and a metal contact that may be coupled to the MTJ stack 100 in the formation of an MRAM memory cell. Depending on the intended composition of the layer, the structural blocking layer 114 may include one or more of Ta, Mo, and W. In one example, during the formation of the structural blocking layer 114, when the structural blocking layer 114 is formed in a processing chamber, the one or more alloys may include at least one of Ta, Mo, and W. The structural blocking layer 114 is a body-centered cubic (bcc) structure oriented in the <100> direction, in contrast to the seed layer 106 and the first and second pinning layers 108 and 112, which can all be oriented in the face-centered cubic <111> direction. When present, the structural blocking layer 114 is from about 0 Å to about 8 Å thick. In one example, the structural blocking layer 114 is formed to a thickness of about 4 Å. In another example, the structural blocking layer 114 is formed directly on and in contact with the second pinning layer 112. In other examples, there is an optional transition layer between the structural barrier layer 114 and the second nailing layer 112.

The magnetic reference layer 116 is formed on the structural barrier layer 114. The magnetic reference layer 116 may be formed by PVD or other suitable process using a single Co-Fe-B alloy, or by using two or more of Co, Fe, and B. In another example, the magnetic reference layer 116 may be formed in a PVD chamber using an alloy target and an element target (such as a CoFe target and a B target) in the presence of Ar plasma. The magnetic reference layer 116 may be formed to a thickness of from about 5 Å to about 10 Å. In one example, the magnetic reference layer 116 may be formed to a thickness of about 10 Å. The magnetic reference layer 116 may include Co _x Fe _y B _z . In one example, x is an integer that, when multiplied by the atomic weight of Co, has a molecular weight percentage of Co _x ranging from about 10% to about 40% of the total molecular weight of the compound (i.e., Co _x Fe _y B _z ); y is an integer that, when multiplied by the atomic weight of Fe, has a molecular weight percentage of Fe _y ranging from about 20% to about 60% of Co _x Fe _y B _z ; and z is an integer that, when multiplied by the atomic weight of B, has a molecular weight percentage of B _z equal to or less than 70% of Co _x Fe _y B _z . The magnetic reference layer 116 may also include different metal combinations with different molecular weight percentages. In another embodiment that may be combined with other embodiments herein, z is an integer that, when multiplied by the atomic weight of B, has a molecular weight percentage of B _z equal to at least 20% of the total compound. In one example, the magnetic reference layer 116 is formed directly on and in contact with the structural barrier layer 114 . In other examples, there is an optional transition layer between the magnetic reference layer 116 and the structural barrier layer 114.

The tunneling barrier layer 118 is formed on the magnetic reference layer 116. The tunneling barrier layer 118 includes a metal oxide, such as magnesium oxide (MgO), hexagonal oxide (HfO ₂ ), titanium oxide (TiO ₂ ), tantalum oxide (TaO _x ), aluminum oxide (Al ₂ O ₃ ) or other materials. In one example, the tunneling barrier layer 118 can be formed in a PVD chamber using a sputtering target of the metal oxide in the presence of Ar. Alternatively, the tunneling barrier layer 118 can be formed using a PVD process in the presence of Ar and O ₂ and a metal of the desired metal oxide. The metal oxide layer can be formed using a PVD process in the presence of O _2. The thickness of the tunneling barrier layer 118 is from about 1 Å to about 15 Å. The exemplary tunnel barrier layer 118 may have a thickness corresponding to the resistance area (RA) requirement of the MTJ stack 100. The resistance area product (RA) is between about 1 Ωμm ² and about 20 Ωμm ² , such as about 10 Ωμm ^2. In one example, the tunnel barrier layer 118 is formed directly on and in contact with the magnetic reference layer 116. In other examples, there is a transition layer between the tunnel barrier layer 118 and the magnetic reference layer 116 that does not affect the performance of the MTJ stack 100.

The MTJ stack 100 also includes a magnetic storage layer 120 formed on the tunneling barrier layer 118. The magnetic storage layer 120 may include one or more layers of Co _x Fe _y B _z . In some examples, the magnetic storage layer 120 may alternatively or additionally include one or more layers of Ta, Mo, W, and Hf. Thus, the deposition of the magnetic storage layer 120 may be performed using a PVD process using Ar plasma and one or more targets made of a Co _x Fe _y B _z alloy, Co, Fe, or B alone, or a combination of alloys and elements such as CoFe and B.

The thickness of the magnetic storage layer 120 may depend on the one or more materials used to form the magnetic storage layer 120. In one example, the magnetic storage layer 120 is made of Co _x Fe _y B _z , where x is an integer such that the molecular weight percentage of Co _x is from about 10% to about 40% of Co _x Fe _y B _z , y is an integer such that the molecular weight percentage of Fe _y is from about 20% to about 60% of Co _x Fe _y B _z , and z is an integer such that the molecular weight percentage of B _z is equal to or less than about 70% of Co _x Fe _y B _z . The thickness of the magnetic storage layer 120 may be from about 5 Å to about 20 Å, and in some examples, the thickness of the magnetic storage layer 120 is about 20 Å. In one example, the magnetic storage layer 120 is formed directly on and in contact with the tunneling barrier layer 118. In other examples, an optional transition layer is disposed between the magnetic storage layer 120 and the tunneling barrier layer 118. The magnetic storage layer is further described below and shown in FIG. 2D.

In the example of MTJ stack 100, capping layer 122 is formed on magnetic storage layer 120. Capping layer 122 includes a plurality of intermediate layers. One or more of the intermediate layers may include an oxide containing Fe. One or more intermediate layers of capping layer 122 may be formed of a dielectric material. In addition, in some examples, hard mask layer 124 is formed directly on and in contact with capping layer 122.

In another example, a hard mask layer 124 is formed on the cap layer 122, and a transition layer is disposed between the cap layer 122 and the hard mask layer 124. The hard mask layer 124 may be formed of a metal oxide, amorphous carbon, a ceramic, a metal material, or a combination thereof. In one example, the magnetic storage layer 120 is formed directly on and in contact with the cap layer 122. In other examples, an optional transition layer is disposed between the magnetic storage layer 120 and the cap layer. The cap layer 122 is shown in FIG. 2E below.

FIG. 2A is an enlargement of the buffer layer 104 of the MTJ stack 100 of FIG. 1. The buffer layer 104 includes tantalum (Ta) or TaN, or a layer stack of Ta and TaN. In some examples, the buffer layer 104 includes Co _x Fe _y B _z alone or in combination with Ta, TaN, or a Ta/TaN layer stack. In an example of the buffer layer 104, the buffer layer 104 includes a double layer 204D. The double layer 204D includes a first buffer middle layer 204A and a second buffer middle layer 204B. Each double layer 204D may include one or more groups of a first buffer interlayer 204A and a second buffer interlayer 204B. In this example, the first buffer interlayer 204A includes Ta, and the second buffer interlayer 204B includes TaN. The first buffer interlayer 204A contacts the substrate 102. In another example, the first buffer interlayer 204A includes TaN, and the second buffer interlayer 204B includes Ta.

In other examples, the buffer layer 104 is made of Co _x Fe _y B _z . Therefore, the Co _x Fe _y B _z material made of the buffer layer 104 is in direct contact with the substrate 102. In another example, as shown in FIG. 2A, a third buffer intermediate layer 204C is formed above at least one double layer 204D. In this example, the third buffer intermediate layer 204C is made of Co _x Fe _y B _z and is formed to a thickness of 10 Å. Therefore, depending on the configuration of the buffer layer 104, the thickness of the buffer layer 104 ranges from 0 Å to about 60 Å thick. In the example of the third buffer intermediate layer 204C, Co _x Fe _y B _z is used, x is an integer such that the molecular weight percentage of Co _x is Co _x Fe _y B _z (i.e., the molecular weight of the compound) is from about 10% to about 40%, y is an integer such that the molecular weight percentage of Fe _y is from about 20% to about 60% of the molecular weight of the compound, and z is an integer such that the molecular weight percentage of B _z is equal to or less than or equal to about 70% of the molecular weight of the compound.

FIG. 2B is an enlarged view of the first pinning layer 108 of the MTJ stack 100 of FIG. 1. In one example, the first pinning layer 108 is made of at least one bilayer 230, and when two or more bilayers are used, the bilayers form a bilayer stack 234. Each bilayer 230 is made of a first intermediate layer 208A and a second intermediate layer 208B. The bilayers 230 of the first pinning layer 108 are represented as (X/Y) _n , (208A/208B) _n , where each bilayer is a combination of X and Y materials, and n is the number of bilayers in the first pinning layer 108. In one example, X is Co, and Y is one of Pt, Ir, Ni, or Pd. Although n=4 in the example shown in FIG. 2B , n is from 3 to 10 in alternative examples.

In an example where the first intermediate layer 208A includes Co and the second intermediate layer 208B includes Pt, the thickness of the first intermediate layer 208A may be from about 1Å to about 7Å, such as from about 1Å to about 3Å. In other examples, the thickness of the second intermediate layer 208B may be from about 1Å to about 8Å, such as from about 1Å to about 3Å. In the example shown in Figure 2B, the first intermediate layer 208A has a thickness of about 2.4Å, and the second intermediate layer 208B has a thickness of about 2.4Å. In another example, the first intermediate layer 208A may have a thickness of about 5Å, and the second intermediate layer 208B may have a thickness of about 3Å. The thickness of the first intermediate layer 208A may be greater than or equal to about 0Å (i.e., no layer exists) to about 10Å. In one example, the thickness of the first intermediate layer 208A is about 5Å.

In an example where the first intermediate layer 208A includes Co and the second intermediate layer 208B includes Ni, the thickness of Co may be from about 1 Å to about 8 Å. In this example, the second intermediate layer 208B may have a thickness from about 1 Å to about 8 Å. The first intermediate layer 208A may have a thickness of about 5 Å. The bilayer 230 is also represented as (X/Y) _n , (208A/208B) _n , where n is 1 to 10 X/Y layer pairs.

In further another example, in addition to the optional capping layer 208C of Co formed on top of at least one bilayer 230, the bilayer 230 is also formed directly on and in contact with the seed layer 106. In examples where the first intermediate layer 208A includes Co and Pt, or Co and Ni, the optional capping layer 208C may be about 0Å to about 10Å thick. Depending on the example, the total thickness of the first pinning layer 108 is from about 0.3nm to about 18nm, and the first pinning layer 108 may include one or more layers including the bilayer 230. In other examples, one or more optional transition layers may be formed between the first pinning layer 108 and the seed layer 106.

FIG. 2C is an enlarged view of the second pinning layer 112 of the MTJ stack 100 of FIG. 1. In one example, the second pinning layer 112 is made of a single cobalt layer. In an alternative example, the second pinning layer 112 is made of at least one bilayer 232. The bilayer 232 includes a first intermediate layer 212A formed of Co and a second intermediate layer 212B formed of one or more of Pt, Ir, Ni, and Pd. When two or more bilayers (such as the bilayer 232) are used in the second pinning layer 112, the plurality of bilayers are referred to as a bilayer stack 236.

In one example, the first intermediate layer 212A is a Co layer from about 1Å to about 7Å thick, and the second intermediate layer 212B is from about 1Å to about 8Å thick. The capping layer 212C of Co may be disposed on and in contact with at least one bilayer 232. In one example, the bilayer 232 is formed directly on and in contact with the coupling layer 110 formed of SyF.

The bilayer 232 of the second pinning layer 112 can be expressed as (X/Y) _n , (212A/212B) _n , where n is the number of bilayers. Although n=4 in the example of Figure 2C, in alternative examples, the bilayer 232 is optional and n can be from 0 to 5. When the first pinning layer 108 also includes a Co/Pt bilayer 230, the bilayer 232 of the second pinning layer 112 can include a Co/Pt stack. In one example of the first intermediate layer 212A, X is Co and Y is Pt. However, Y can also be Ir, Ni, or Pd. Although n=4 in the example of Figure 2C, in alternative examples, n is from 0 to 5. In the case where the first intermediate layer 212A includes Co and the second intermediate layer 212B includes Pt, the thickness of the first intermediate layer 212A may be from about 1Å to about 3Å, or the thickness of the first intermediate layer 212A may be from about 3Å to 7Å. In other examples, the second intermediate layer 212B may have a thickness from about 1Å to about 3Å, or the second intermediate layer 212B may have a thickness from about 3Å to about 8Å. In the example shown in FIG. 2C, the first intermediate layer 212A has a thickness of about 2.4Å, and the second intermediate layer 212B has a thickness of about 2.4Å. In another example, the first intermediate layer 212A may have a thickness of about 3Å, and the second intermediate layer 212B may have a thickness of about 3Å.

In another example, when the first pinning layer 108 includes a Co/Ni bilayer 230, the bilayer 232 of the second pinning layer 112 may include a Co/Pt stack. In this example, the first intermediate layer 212A includes Co and the second intermediate layer 212B includes Pt. The first intermediate layer 212A may have a thickness from about 0.5Å to about 8Å. The second intermediate layer 212B may have a thickness from about 0.5Å to about 7Å. The number of layers "n" in the bilayer 232 may be between 0 and 5.

The Co capping layer 212C is disposed on top of and in contact with at least one bilayer 232. In some examples, an optional transition layer may be used between the bilayer 232 and the second pinning layer 112 and/or between the bilayer 232 and the SyF coupling layer 110 (or both).

In both the Co/Pt and Co/Ni examples discussed above, the capping layer 212C may have a thickness of up to about 10Å. In one example, where the capping layer 212C includes Co, the thickness is about 5Å. The total thickness of the second pinning layer 112 is from about 0.3nm to about 18nm, and the second pinning layer 112 may include one or more layers including the bilayer 232 discussed herein.

In one example, the first nailing layer 108 and the second nailing layer 112 may each have substantially the same composition and/or the same thickness. In an alternative example, the first nailing layer 108 and the second nailing layer 112 may each have different compositions and/or thicknesses. In one example, the first nailing layer 108 is Co, and the second nailing layer 112 is formed of a bilayer 232. Each bilayer of the second nailing layer 112 may include a first intermediate layer 212A formed of Co and a second intermediate layer 212B formed of Pt. In this way, one or more Co/Pt bilayers are formed. In another example, the first pinning layer 108 is formed of Co, and the second pinning layer 112 is formed of one or more Co/Ni bilayers. In yet another example, the first pinning layer 108 includes one or more bilayers, each bilayer containing a first intermediate layer of Co and a second intermediate layer of Ni, and the second pinning layer 112 includes one or more bilayers, each bilayer containing a first intermediate layer of Co and a second intermediate layer of Pt. In yet another example, the first pinning layer 108 includes one or more bilayers, each bilayer including a first intermediate layer of Co and a second intermediate layer of Pt, and the second pinning layer 112 includes one or more bilayers, each bilayer including a first intermediate layer of Co and a second intermediate layer of Ni.

FIG. 2D is an enlarged view of an exemplary magnetic storage layer 120 of the MTJ stack 100 of FIG. 1. The magnetic storage layer 120 is made of three layers, namely, a first magnetic layer 220A, a second magnetic layer 220B, and a non-magnetic layer 220C disposed between the first magnetic layer 220A and the second magnetic layer 220B. The first magnetic layer 220A of the magnetic storage layer 120 and the second magnetic layer 220B of the magnetic storage layer 120 are both made of Co _x Fe _y B _z . In one example, x is an integer such that the molecular weight percentage of Co _x is from about 10% to about 40% of the molecular weight of Co _x Fe _y B _z (i.e., the molecular weight of the compound), y is an integer such that the molecular weight percentage of Fe _y is from about 20% to about 60% of the molecular weight of the compound, and z is an integer such that the molecular weight percentage of B _z is less than or equal to about 70% of the molecular weight of the compound. The non-magnetic layer 220C is made of at least one of Ta, Mo, W, and Hf or one or more combinations thereof. The non-magnetic layer 220C may contain dopants (such as boron, oxygen, or other dopants). The non-magnetic layer 220C enhances the pinning moment perpendicular to the substrate plane (e.g., in the y-direction 170 that is consistent with the stacking direction of the MTJ stack 100 and perpendicular to the plane of the substrate 102), which promotes magnetic anisotropy, that is, the directional dependence of the magnetic properties of the structure. The first magnetic layer 220A may have a thickness between about 5Å and about 20Å. The thickness of the second magnetic layer 220B is between about 5Å and about 20Å. Therefore, the thickness of the non-magnetic layer 220C may be between about 0Å and about 8Å. In another example, the thickness of the first magnetic layer 220A is between about 8Å and about 10Å. The thickness of the second magnetic layer 220B is between about 6Å and about 12Å, and the thickness of the non-magnetic layer 220C is between about 1Å and about 2Å.

FIG. 2E is an enlarged view of an exemplary capping layer 122 of the MTJ stack 100 of FIG. 1 . The total thickness of the capping layer 122 is between about 2 Å and about 110 Å. In one example, the total thickness of the capping layer including all the intermediate layers is about 60 Å. It should be understood that the capping layer 122 may include a plurality of intermediate layers. For example, the capping layer 122 may have a first capping intermediate layer 222A, a second capping intermediate layer 222B, a third capping intermediate layer 222C, a fourth capping intermediate layer 222D, or even more capping intermediate layers. The first cover intermediate layer 222A is made of MgO or another iron-containing oxide formed directly on the magnetic storage layer 120, and its thickness is from about 2Å to about 10Å. On top of the first cover intermediate layer 222A, a second cover intermediate layer 222B of at least one or one or more combinations of Co, Fe and B is formed, and its thickness is between about 0Å and about 20Å. In one example, the thickness of the second cover intermediate layer 222B is between 6Å and 8Å. The third cover intermediate layer 222C is optionally formed of at least one or one or more combinations of Mo, Ta and W, and is formed on the second cover intermediate layer 222B, and its thickness is between about 0Å and about 30Å. The third cover interlayer 222C may have a thickness between about 8Å and about 12Å. When molybdenum is included in the third cover interlayer 222C, a larger lattice is produced. The larger lattice in the third cover interlayer 222C increases the tensile stress on the MgO in the first cover interlayer 222A. The increase in tensile stress in the dielectric elements of the cover layer 122 improves the perpendicular magnetic anisotropy of the metal elements in the cover layer 122. It should be understood that some examples of the cover layer 122 do not contain the third cover interlayer 222C. In one or more examples, the fourth cover interlayer 222D is optionally formed on the third cover interlayer 222C. The fourth cover interlayer 222D is formed of at least one or more of Ru, Ir and a combination thereof, and has a thickness between about 0 Å and about 50 Å. In one example, the thickness of the fourth cover interlayer 222D is between about 20 Å and about 30 Å. It should be understood that the cover layer 122 includes the first cover interlayer 222A, and may include one or more of the second cover interlayer 222B, the third cover interlayer 222C, the fourth cover interlayer 222D or other cover interlayers. An optional transition layer may be disposed between some or all of the first covering interlayer 222A, the second covering interlayer 222B, the third covering interlayer 222C, the fourth covering interlayer 222D, or other covering interlayers. In addition, an optional transition layer may be disposed between the covering layer 122 and the magnetic storage layer 120.

FIG. 3 shows a hysteresis curve 300 of the magnetic coherence of the magnetic storage layer 120 shown in FIG. 1. The hysteresis curve illustrates a phenomenon in which the value of a physical property lags behind the effect causing the change in the physical property. Here, the magnetic moment lags behind the changing magnetizing force and is directly related to the magnetic coherence of the magnetic storage layer 120. The magnetic coherence is a measure of the ability of a ferromagnetic material to withstand an external magnetic field without being demagnetized. To determine the magnetic coherence, the magnetic moment ( ^A.m2 ) is measured relative to the changing magnetic force, applied in units of kilo-Oersted (kOe). The absolute value of the magnetic coherence 304 of the magnetic storage layer of the MTJ stack 100 has been determined to be greater than 0.15 kOe. The level of magnetic coherence 304 for the MTJ stack 100 can be increased by increasing the coupling between the first magnetic layer 220A and the second magnetic layer 220B. Thinning the nonmagnetic layer 220C additionally increases the coupling between the first magnetic layer 220A and the second magnetic layer 220B, and thus additionally increases the level of magnetic coherence 304 for the MTJ stack 100.

FIG. 4 is an enlarged view of an exemplary first magnetic layer 220A of the magnetic storage layer 220 of FIG. 2D. It should be understood that the following discussion is not limited to the first magnetic layer 220A, but is applicable to any magnetic layer discussed herein, including the second magnetic layer 220B. By increasing the coupling between the first magnetic layer 220A and the second magnetic layer 220B, the thickness 408 of the magnetic dead layer 400 is reduced. The magnetic dead layer 400 is a sublayer of the first magnetic layer 220A having magnetic material but no magnetic moment. The dimensions (150, 170 and z-direction 160) of the first magnetic layer 220A having the first magnetic layer 220A and the second magnetic layer 220B may be in the nanometer or Å range. The lateral dimensions (150 and 160) and thickness 412 of the first magnetic layer 220A together define a magnetic volume 404. The magnetic dead layer 400 does not contribute to the magnetic volume 404, and therefore reduces the magnetization of the first magnetic layer 220A in the entire thickness direction of the first magnetic layer 220A (i.e., the y-direction 170). One potential cause of the magnetic dead layer 400 is due to the intermixing between the magnetic material and other non-magnetic materials. Therefore, reducing the intermixing between the magnetic materials (Fe, Co) of the first magnetic storage layer 220A and the second magnetic storage layer 220B and the transition metals (Ta, W, Mo, Hf) of the non-magnetic layer 220C also reduces the thickness and lateral dimensions of the magnetic dead layer 400. The non-magnetic layer 220C is substantially non-metallic. As previously mentioned, the magnetic storage layer 120 of the MTJ stack 100 disclosed herein can retain data for up to 10 years at 125 degrees Celsius.

In the known MTJ stack structure with the known magnetic storage layer (shown by curve 998), the absolute value of the known magnetic hysteresis 999 of the known magnetic storage layer is about 0.05 kOe. The known MTJ stack structure (curve 998) is shown on the same graph, offset by 0.4 to highlight the difference from the hysteresis curve 300. The known magnetic storage layer does not have the same structure as the magnetic storage layer 120 disclosed herein, and therefore has a more known magnetic hysteresis 999.

Advantageously, the magnetic refractive force 304 of the magnetic storage layer 120 is more than twice the known magnetic refractive force 999 of the known MTJ stack structure (i.e., curve 998). In addition, the MTJ stack 100 of the present disclosure exhibits a tunneling magnetoresistance (TMR) of more than 150% at a resistance area product (RA) of ^10Ωμm2 . The magnetic storage layer 120 also improves the bake error rate. The bake error rate is the error bit rate (bit flip) after baking the various layers included in the magnetic storage layer 220, as shown in FIG. 3.

Examples of MTJ stacks for STT MRAM memory are disclosed herein. Although the foregoing relates to examples of this disclosure, other and further examples of this disclosure may be devised without departing from the basic scope of this disclosure, and the scope of this disclosure is determined by the scope of the following patent applications.

110: Coupling layer/ferromagnetic (SyF) coupling layer

212: Second nailing layer

212A: First middle layer

212B: Second middle layer

212C: Covering layer

232: Double layer

236: Double stacking

Claims (17)

A magnetic tunneling junction stack includes: a structural blocking layer; a magnetic reference layer disposed on the structural blocking layer, the magnetic reference layer being formed of cobalt (Co) or iron (Fe); a tunneling barrier layer disposed on the magnetic reference layer, the tunneling barrier layer being formed of magnesium (Mg); and a magnetic storage layer disposed in contact with the tunneling barrier layer, the magnetic storage layer including: a first magnetic layer, the first magnetic layer being formed of cobalt (Co), iron (Fe) and boron ( B); a non-magnetic layer formed of one of molybdenum (Mo), tantalum (Ta), tungsten (W) or a combination thereof; and a second magnetic layer formed of Co, Fe and B, the non-magnetic layer being arranged to contact the first magnetic layer and the second magnetic layer, wherein a thickness of the non-magnetic layer is less than a thickness of the first magnetic layer, and the thickness of the non-magnetic layer is proportional to a thickness of a magnetic dead layer in the magnetic storage layer. A magnetic tunneling junction stack as described in claim 1, wherein the non-magnetic layer is less than or equal to 2Å thick. A magnetic tunneling junction stack as described in claim 1, wherein the magnetic storage layer is less than or equal to 20Å thick. The magnetic tunneling junction stack as described in claim 1, comprising: a buffer layer in contact with a seed layer, wherein the buffer layer is formed of cobalt (Co), tantalum (Ta) or a combination thereof; a SyF coupling layer disposed on the seed layer, wherein the SyF coupling layer is formed of ruthenium (Ru) or iridium (Ir); and a structural blocking layer disposed on the SyF coupling layer, wherein the structural blocking layer includes molybdenum (Mo), tantalum (Ta) or tungsten (W). The magnetic tunneling junction stack as described in claim 4 further comprises: a first pinning layer disposed on the seed layer and in contact with the SyF coupling layer, a second pinning layer disposed on the SyF coupling layer and in contact with the structural barrier layer, wherein the second pinning layer is formed by a transition metal. The magnetic tunneling junction stack as described in claim 1 further comprises: a capping layer comprising an oxide of magnesium or iron. The magnetic tunneling junction stack as described in claim 1 further comprises: a capping layer, the capping layer comprising: a first capping intermediate layer comprising an oxide of magnesium or iron; and a second capping intermediate layer comprising at least one of Co, Fe and B. A magnetic tunneling junction stack includes: a first pinning layer; a magnetic reference layer disposed on the first pinning layer, the magnetic reference layer being formed of cobalt (Co); a tunneling barrier layer disposed in contact with the magnetic reference layer; and a magnetic storage layer disposed in contact with the tunneling barrier layer, the magnetic storage layer including: a first magnetic layer disposed in contact with a non-magnetic layer, the first magnetic layer being formed of Co, iron (Fe) and boron (B); the non-magnetic layer , formed of a transition metal, wherein the transition metal of the non-magnetic layer is made of at least one of the following: molybdenum (Mo), tantalum (Ta) or tungsten (W); and a second magnetic layer, formed of Co, Fe and B, the non-magnetic layer is arranged to contact the first magnetic layer and the second magnetic layer, wherein a thickness of the non-magnetic layer is less than a thickness of the first magnetic layer, and the thickness of the non-magnetic layer is proportional to a thickness of a magnetic dead layer in the magnetic storage layer. A magnetic tunneling junction stack as described in claim 8, wherein the non-magnetic layer is less than or equal to 2Å thick. A magnetic tunneling junction stack as described in claim 8, wherein the magnetic storage layer is less than or equal to 20Å thick. The magnetic tunneling junction stack as described in claim 8 further comprises: a capping layer, the capping layer comprising: a first capping intermediate layer comprising an oxide of magnesium or iron; and a second capping intermediate layer comprising at least one of Co, Fe and B. A magnetic tunneling junction stack as described in claim 8, wherein the first magnetic layer can be represented by Co _x Fe _y B _z , wherein a molecular weight of Co is greater than or equal to approximately 10 weight % and less than or equal to approximately 40 weight % of the first magnetic layer, and a molecular weight of Fe is greater than or equal to approximately 20 weight % and less than or equal to approximately 60 weight % of the first magnetic layer. A magnetic tunneling junction stack includes: a magnetic reference layer; a tunneling barrier layer, the magnetic reference layer is disposed on the tunneling barrier layer, the magnetic reference layer is formed of cobalt (Co); a magnetic storage layer is disposed to contact the tunneling barrier layer, wherein the magnetic storage layer includes: a first magnetic layer, contacting a non-magnetic layer, the first magnetic layer is formed of Co, iron (Fe) and boron (B); the non-magnetic layer is formed of molybdenum (Mo), tantalum (Ta) and tungsten (W) or a combination thereof; and a second magnetic layer, formed of Co, Fe and B, the non-magnetic layer separates the first magnetic layer and the second magnetic layer, wherein a thickness of the non-magnetic layer is less than that of the first magnetic layer. The non-magnetic layer and the second magnetic layer have a thickness, and the thickness of the non-magnetic layer is configured to correspond to the thickness of a magnetic dead layer in the magnetic storage layer; and a covering layer, arranged to contact the magnetic storage layer, wherein the covering layer includes: a first covering middle layer; a second covering middle layer; and a third covering middle layer, the second covering The intermediate layer is formed of at least one of the same elements as the first magnetic layer, the third covering intermediate layer is formed of at least one of Mo, Ta or W, the first covering intermediate layer is formed of an oxide of an alkaline earth metal or an oxide of a transition metal, and the second covering intermediate layer separates the first covering intermediate layer and the third covering intermediate layer. A magnetic tunneling junction stack as described in claim 13, wherein the non-magnetic layer is less than or equal to 2Å thick. A magnetic tunneling junction stack as described in claim 13, wherein the magnetic storage layer is less than or equal to 20Å thick. A magnetic tunneling junction stack as described in claim 13, wherein: the first covering interlayer comprises an oxide of magnesium or iron; and the second covering interlayer comprises at least one of Co, Fe and B. A magnetic tunneling junction stack as described in claim 13, wherein the first magnetic layer can be represented by Co _x Fe _y B _z , wherein the molecular weight of Co is greater than or equal to approximately 10 wt % and less than or equal to approximately 40 wt % of the first magnetic layer, and the molecular weight of Fe is greater than or equal to approximately 20 wt % and less than or equal to approximately 60 wt % of the first magnetic layer.

Images