Method for etching a magnetic tunnel junction layer stack

Abstract

A method for etching a magnetic tunnel transition layer stack, hereinafter referred to as an MTJ layer stack, comprising: (a) providing an MTJ layer stack on a first electrode (10), wherein the MTJ layer stack comprises a top hard mask layer (15) and a first layer stack with a reference layer, a free layer, and a tunnel barrier layer between the reference layer and the free layer; (b) forming a structure in the hard mask layer by an etching step (31) which is an inert gas ion beam etching or a reactive ion etching with a fluorocarbon or carbon chloride, wherein the structure has a sidewall extending from a hard mask top to a top of the first layer stack;and (c) forming a structure in the first layer stack with a sidewall (20) forming a continuous surface with the sidewall in the hard mask layer and extending to an upper surface of the first electrode (10), wherein the structure in the first layer stack is generated by a process sequence comprising: (1) a first etching step (32i) which is an IBE with an inert gas; and (2) a second etching step (32r) which is a single reactive ion etching step with one or more chemicals selected from ethanol, ammonia and carbon monoxide, wherein the second etching step (32r) oxidizes non-volatile residues (19) on the sidewall of the first layer stack generated by the first etching step (32i) to obtain volatile residues (19x).

Description

RELATED PATENT APPLICATION

[0001] This application relates to US patent 8,722,543, which has been transferred to the same applicant and is incorporated herein in its entirety by reference. TECHNICAL AREA

[0002] The present disclosure relates to a method for reducing MTJ sidewall damage during an etching process, which transfers a mask structure through an MTJ layer stack, thereby resulting in an array of MTJ nanopillars with an improved magnetoresistive ratio and other magnetic properties. TECHNICAL BACKGROUND

[0003] An MTJ storage element, also known as an MTJ nanopillar, is a key component in magnetic recording devices and storage devices such as magnetoresistive random access memory (MRAM) and spin torque transfer MRAM (STT-MRAM). A crucial step in fabricating an array of MTJs is the etch transfer of a structure in an overlying hard mask through an MTJ layer stack to form an array of MTJ nanopillars with a critical dimension (CD), which in conventional devices is significantly less than 100 nm in plan view. The etch transfer process typically involves multiple etching steps using reactive ion etch (RIE) and / or ion beam etch (IBE).

[0004] An MTJ layer stack comprises two ferromagnetic layers, designated as the free layer (FL) and the reference layer (RL), and a dielectric layer (tunnel barrier) between the FL and the RL. The RL has a fixed magnetization, preferably in a direction perpendicular to the plane (referred to as perpendicular magnetic anisotropy, or PMA), while the FL can rotate freely in a direction parallel or antiparallel to the RL magnetization direction, thus providing a "0" or "1" memory state for the MTJ. The magnetoresistive ratio is expressed by dR / R, where dR is the resistance difference between the two magnetic states when a current is passed through the MTJ, and R is the minimum resistance value.

[0005] The bottom MTJ layer is typically a non-magnetic seed layer that promotes uniform growth of the overlying layers and enhances the PMA in the overlying RL or FL. A cover layer, such as Ta, is generally formed as the top MTJ layer and serves as a protective layer during subsequent physical and chemical etching processes. Thus, a single etch transfer process through the MTJ layer stack presents a challenge because different materials (magnetic alloys, non-magnetic metals, and dielectric layers) exhibit varying etch rates when subjected to IBE with Ar or a conventional CH3OH-based RIE. In particular, a methanol RIE causes chemical and plasma damage to MTJ sidewalls, although there is minimal redeposition of etched material on the sidewalls.On the other hand, IBE does not cause chemical damage and leaves minimal plasma damage, but it does result in a high proportion of redeposited material on MTJ sidewalls. If a metal like Ta is redeposited on the tunnel barrier, short circuits can easily occur, rendering the device unusable.

[0006] Conventional technology offers no solution for etching to transfer a hard mask structure across an entire MTJ layer stack without significant redeposition of one or more MTJ materials on the MTJ sidewalls, nor without significant damage to the sidewalls. In either case, material removal from the sidewalls requires one or more additional steps, reducing throughput and increasing costs. Furthermore, damaged sidewalls are difficult to repair and often result in reduced yield and thus higher costs per unit of an acceptable product. Therefore, a novel method for etching an MTJ layer stack in a single etching process is needed to achieve higher throughput and lower costs, and the method should maintain or even improve the magnetic properties, including the magnetoresistive ratio, in the MTJ nanopillar.Furthermore, a process flow for etching MTJ sidewalls is desired that significantly reduces sidewall damage in devices with a diameter (CD) of approximately 60 nm or less. SUMMARY

[0007] One objective of the present disclosure is to provide a method for etching all layers in an MTJ stack below the hard mask with a single etching step that produces minimal residue.

[0008] A second objective of the present disclosure is to provide a process flow for MTJ etching that fulfills the first objective and significantly reduces sidewall damage and the associated edge effects compared to conventional methanol-based RIE, thereby enabling improved performance, especially for MTJ nanocolumns with critical dimensions ≤ 60 nm.

[0009] According to a preferred embodiment, the first objective is achieved with an MTJ layer stack comprising at least one reference layer, a free layer, a tunnel barrier between the free layer and the reference layer, and a top hard mask. In some embodiments, a seed layer is used as the bottom MTJ layer. A structure comprising multiple island features with the desired critical dimension for the subsequent MTJ nanopillars is initially defined in a photoresist mask layer above the hard mask layer. Preferably, a bottom antireflective coating (BARC) or a dielectric antireflective coating (DARC) is located between the hard mask and the photoresist mask layer, exhibiting better resistance to subsequent etching processes than the photoresist mask.The structure is transferred by means of a first RIE or IBE step through the BARC or the DARC and then transferred through the hard mask by continuing the first etching step, or by performing a second RIE step which includes a fluorocarbon gas or chlorocarbon gas, or by means of a second IBE step.

[0010] According to one embodiment of the present disclosure, the island feature structure is then transferred through the remaining MTJ layers by means of an RIE step comprising an inert gas and a chemical such as methanol, ethanol, ammonia, or a combination of CO and NH3. In particular, Ar ions or the like constitute a physical component for etching, while one or more chemicals constitute a plasma component for chemical etching. As a result, inert gas ions or plasma significantly minimize chemical damage to the MTJ sidewalls, and the chemical component substantially reduces the redeposition of etched residues on the MTJ sidewalls. In fact, depending on the composition of the MTJ layers, the etching conditions can be optimized to achieve minimal sidewall residues and significantly less sidewall damage than with conventional methanol-based RIE.An encapsulation layer is then deposited onto the resulting MTJ nanopillars, followed by a chemical mechanical polish (CMP) process to remove any layers above the hard mask. The CMP process creates a hard mask surface that is coplanar with the surrounding encapsulation layer. In a top view, the MTJ nanopillars form, for example, an array of circular or elliptical shapes.

[0011] In a second embodiment, a process sequence is used to transfer the island shape structure in the hard mask through the remaining MTJ layers, thereby generating multiple MTJ nanopillars exhibiting minimal sidewall damage and residue. A first process flow includes an IBE step followed by a separate chemical treatment to convert non-volatile residues on MTJ sidewalls into volatile residues. Subsequently, a second IBE step, a plasma sputtering etching step, or a thermal treatment is performed to remove the volatile residue. A second process flow includes an RIE step followed by a separate chemical treatment, where the chemical can be applied without plasma, to convert a sidewall residue into a volatile shape. Next, an optional IBE step, a plasma sputtering etching step, or a thermal treatment is used to remove the volatile residue.In other embodiments, RIE is alternated with IBE prior to the optional chemical treatment and with an optional volatilization step. List of characters Fig. Figure 1 is a cross-sectional view of an MTJ layer stack on which a photoresist structure is formed and shows ions by which the structure is transferred through an underlying BARC or DARC during an etching sequence of the present disclosure. Fig. Figure 2 is a cross-sectional view of an MTJ stack. Fig. 1 according to an embodiment of the present disclosure, after an etching process has transferred the structure through the top hard mask MTJ layer. Fig. Figure 3 is a cross-sectional view of an MTJ nanopillar after etching according to an embodiment described herein, in which physical and chemical components are combined to create the hard mask structure of the Fig. 2 to transfer through the remaining MTJ layer stack. Fig. Figure 4a is a cross-sectional view of the MTJ stack of the Fig. 2, after ion beam etching transfers the hard mask structure through the MTJ layer stack and leaves residues on MTJ sidewalls. Fig. Figure 4b is a cross-sectional view of the MTJ stack of the Fig. 2, after reactive ion etching transfers the hard mask structure through the MTJ stack and leaves residues on MTJ sidewalls. Fig. Figure 5 is a cross-sectional view of the MTJ nanopillar of the Fig. 4a or Fig. 4b after a chemical treatment according to an embodiment of the present disclosure, in which non-volatile residues are converted to volatile residues on MTJ sidewalls. Fig. Figure 6 is a cross-sectional view of the MTJ nanopillar in Fig. 5 after a volatilization step with IBE, plasma sputtering etching or heat treatment according to an embodiment of the present disclosure to remove the volatile residue. Fig. Figure 7 is a cross-sectional view of the MTJ nanopillar in Fig. 6 after deposition of an encapsulation layer and planarization for electrical isolation of the MTJ nanopillar from neighboring MTJ nanopillars. Fig. Figure 8 is a top view of several MTJ nanopillars having a circular shape in an arrangement of rows and columns according to an embodiment of the present disclosure. Fig. Figure 9 is a flowchart illustrating a sequence of steps for forming an MTJ nanopillar according to an embodiment of the present disclosure. Fig. 10 and Fig. Figure 11 are flowcharts showing alternative sequences for the formation of MTJ nanopillars with sidewalls that are substantially free of damage and residue according to the embodiments of the present disclosure. Fig. Figure 12 is a diagram of the magnetoresistive ratio plotted against MTJ size for MTJs etched without a post-cleaning process, while Fig. 13 a similar diagram for MTJ nanopillars formed by a combined physicochemical etching followed by an Ar plasma etching volatilization step according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0012] The present disclosure relates to a method for etching an MTJ layer stack in which all layers below the hard mask are removed by a single etching process comprising both physical and chemical components to form an MTJ nanopillar with sidewalls exhibiting significantly less sidewall damage and minimal residue compared to conventional methanol-based RIE. An alternative process sequence is provided in which the physical and chemical etchants alternate or are applied in separate steps, followed by a chemical treatment and an optional volatilization process to achieve improved MTJ performance. Although only one MTJ nanopillar is shown in cross-sectional views in the drawings, a person skilled in the art will recognize that multiple MTJ nanopillars can be formed in a typical storage device structure.A process is defined as a procedure comprising one or more steps, and a sequence or process flow according to the present disclosure refers to two or more processes in succession.

[0013] With reference to Fig. 1 becomes an MTJ layer stack 1 , which ultimately becomes several MTJ nanopillars in an MRAM or STT-MRAM, is shown in a cross-sectional view. The y-axis is perpendicular to the planes of the layers in the MTJ stack. There is a substrate 10 , which in one embodiment is a lower electrode in a storage device. The lower electrode can be a multilayer structure and is typically embedded in a dielectric layer (not shown).

[0014] The MTJ stack 1 is on the substrate 10 laid down and in the exemplary embodiment has a lower spin valve configuration, wherein a germ layer 11, pinned or reference layer 12 , tunnel barrier 13 , free shift 14 and hard mask 15 The reference layers and the free layer are formed sequentially on the substrate. Each of the reference layers and the free layer preferably has a respective PMA with magnetization aligned in the y-axis direction. In other embodiments, at least one additional layer can be included in the aforementioned MTJ stack, such as an Hk reinforcement layer between the free layer and the hard mask, which enhances the PMA in the free layer. The nucleation layer can comprise one or more materials such as NiCr, Ta, Ru, Ti, TaN, Cu, Mg, or other materials typically used to promote a smooth and uniform grain structure in overlying layers.

[0015] The reference layer 12The system may have a synthetic antiparallel (SyAP) configuration, represented by AP2 / Ru / AP1, wherein an antiferromagnetic coupling layer of Ru, Rh, or Ir, for example, is located between an AP2 magnetic layer and an APi magnetic layer (not shown). The AP2 layer, also referred to as the outer pinned layer, is formed on the seed layer, while AP1 is the inner pinned layer and is typically in contact with the tunnel barrier. The AP1 and AP2 layers may comprise CoFe, CoFeB, Co, or a combination thereof. In other embodiments, the reference layer may be a laminated stack with an inherent PMA, such as (Co / Ni). n , (CoFe / Ni) n , (Co / NiFe)n, (Co / NiFe) n , (Co / Pt) n , (Co / Pd) nor the like, where n is the number of laminations. Furthermore, a transition layer such as CoFeB or Co can be inserted between the top layer in the laminated stack and the tunnel barrier layer.

[0016] The tunnel barrier layer 13 The metal oxide is preferably one of MgO, TiOx, AlTiO, MgZnO, Al2O3, ZnO, ZrOx, HfOx, or MgTaO. Preferably, MgO is chosen as the tunnel barrier layer because it offers the highest magnetoresistive ratio, especially when, for example, it is located between two CoFeB layers.

[0017] The free shift 14The free layer can be Co, Fe, CoFe, or an alloy thereof with one or both of B and Ni, or it can be a multilayer stack comprising a combination of the aforementioned compositions. In another embodiment, the free layer can include a non-magnetic moment-reducing layer, such as Ta or Mg, inserted between two ferromagnetically coupled CoFe or CoFeB layers. In an alternative embodiment, the free layer has a SyAP configuration such as FLi / Ru / FL2, where FL1 and FL2 are two magnetic layers that are antiferromagnetically coupled, or it is a laminated stack with an inherent PMA, as described above with reference to the composition of the reference layer.

[0018] The hard mask 15Also referred to as the cover layer, it typically comprises one or more materials including Ta, Ru, TaN, Ti, Ti, TiN, and W. It should be noted that other hard mask materials, including MnPt, can be selected to ensure high etch selectivity with respect to the underlying MTJ layers during an etching process that forms MTJ nanopillars with sidewalls that terminate at the bottom electrode. All layers in the MTJ stack can be deposited in a DC sputtering chamber of a sputtering system such as an Anelva C-7100 sputtering system, which includes ultra-high vacuum DC magnetron sputtering chambers with multiple targets and at least one oxidation chamber. The sputtering process typically involves an argon sputtering gas and a base pressure between 5 × 10⁻⁶ mN. -8 and 5 × 10 -9 Torr.

[0019] Once all layers 11-15 Once the MTJ stack is in place, it can be removed. 1The layers are annealed by heating to a temperature between approximately 360 °C and 400 °C for up to several hours to form a bcc structure in the reference layer, the free layer, and the tunnel barrier layer, thereby enhancing PMA in the reference layer and the free layer. It is assumed that the appropriate crystal structure in the aforementioned layers further improves the magnetoresistive ratio in the resulting MTJ nanopillars.

[0020] As a first step in the MTJ structuring process according to the present disclosure, a BARC layer or a DARC layer is used. 16 and a photoresist layer 17 one after the other on the top side 15t applied to the hard mask. The BARC or DARC with a surface 16tIt has a refractive index that minimizes light reflection during subsequent patternwise exposure, thus enabling the formation of more uniform island shapes with less CD variation in the photoresist layer. A conventional patternwise exposure and development sequence is then used to form a structure in the photoresist layer that includes multiple islands with sidewalls. 20 includes. As shown in the top view below of the Fig. As shown in Figure 8, the islands are arranged in a multi-row, multi-column configuration. Fig. Figure 1 shows only a single island for simplicity. Each island has a critical dimension w1, which in some embodiments is between 60 nm and 100 nm, and in others between approximately 10 nm and 60 nm, meeting the requirements for CDs in industry-standard storage devices. It should be noted that some devices are circular, so w1 is formed in both the x- and z-directions. The top-down shape of the island 17 However, it can be an ellipse or a polygon, so that the dimension of the z-axis differs from the dimension of the x-axis.

[0021] In the first etching step 30 , which may be an IBE with one or more Ar, Kr, Xe or Ne or may include a RIE with a fluorocarbon gas or chlorocarbon gas, will form the island shape in the photoresist layer 17 through the BARC layer or DARC layer 16transferred. Accordingly, the side wall extends 20 now from a top side of the photoresist layer to a top side 15t the hard mask 15 , and CD w1 is duplicated in the DARC or BARC layer. The photoresist layer can then be removed using a standard method or will be etched away in subsequent etching processes.

[0022] With reference to Fig. 2. A second etching process will take place. 31 carried out to create the island shape with side wall 20 and CD w1 through the hard mask 15 to transfer. In some embodiments, a second IBE with an inert gas or a second RIE based on fluorocarbons or carbon chlorides can be used. Furthermore, the second IBE or RIE can comprise oxygen. In other embodiments, however, the present disclosure assumes that the etching process is carried out using a different type of RI. 31The conditions applied are essentially the same as in the etching process. 30 , so that the etch transfer through the hard mask is a continuation of the etch transfer through the DARC or BARC layer 16 can be. In other words, etching can 30 the Fig. 1 until the break on the top side 14t the free layer will continue. As mentioned previously, it is likely that the etch transfer process through the hard mask will affect the remaining photoresist layer. 17 removed, since the etch rate through it is usually high relative to the hard mask, and the hard mask can be significantly thicker than the remaining photoresist layer once the etching process is complete. 30 the exposed DARC or BARC layer 16exposes. In some embodiments, a passivation step is performed which includes the introduction of an oxygen plasma or an oxygen flow into the process chamber immediately after completion of the second RIE or IBE and without interrupting a vacuum, in order to produce a smoother sidewall. 20 to produce.

[0023] In the embodiments described herein, it is understood that the IBE typically comprises rotating the workpiece (wafer) on which the MTJ layer stack is formed. Furthermore, the angle of incidence or penetration of noble gas ions directed towards the wafer surface can be between 0° and 90°. The IBE can be employed in one or more of the following steps described below: hard mask etching, MTJ etching, cleaning, and volatilization. On the other hand, the RIE is used only for hard mask or MTJ etching, employs a chemical reactant and a stationary wafer, and the resulting plasma is restricted to a direction of 90° or orthogonal to the wafer surface. According to the present disclosure, a plasma sputtering etching process is used only for volatilization or cleaning steps, comprises a noble gas, and is also restricted to a direction of 90° (orthogonal to the wafer surface).

[0024] With reference to Fig. 3 relates to a key feature according to a first embodiment of the present disclosure, a single RIE step 32m , which effectively creates the island shape in the hard mask through all underlying MTJ layers 11 until 14 transmits and thus the MTJ nanopillar 1a forms. The etching step 32mThis process comprises a physical component, represented by inert gas ions or a plasma, and a chemical component, comprising ions or a plasma of chemical species generated under reactive ion etching conditions. Herein, the inert gas is defined as a noble gas, specifically one of Ar, Kr, Ne, or Xe. Thus, both an inert gas and one or more chemicals, including methanol, ethanol, H₂O₂, H₂O, N₂O, NH₃, and CO, are introduced into an etching chamber, while a plasma is ignited with a radio frequency (RF) power between 600 watts and 3000 watts at a temperature near room temperature. It is understood that the RF power received at an upper electrode in a RIE chamber may differ from that received at a lower electrode. Typically, one or more wafers are held at the lower electrode during the RIE process.In a preferred embodiment, the resulting ions and plasma are aligned orthogonally to the top surface of the substrate along the y-axis direction. Consequently, the side wall is... 20 a continuous surface extending from a top surface 15t the hard mask 15 up to the upper surface 10t the lower electrode extends.

[0025] In a preferred embodiment, the side wall 20 essentially vertically, so that the CD w1 is in all MTJ layers 11-15 essentially remains unchanged. It should be noted that the sidewall angle α tends to become more vertical as the methanol content (or chemical content) in the noble gas / chemical mixture of the step increases. 32mThe angle α increases. For example, it may be close to 75° when the noble gas content is 100%, but becomes essentially 90° at a chemical content of approximately 50% or more. The term "content" here refers to the flow rate ratio. Thus, a 50:50 noble gas to chemical flow rate ratio indicates a chemical content of 50% and a noble gas content of 50% in the RIE gas mixture. Furthermore, we surprisingly found that the combined physical and chemical etching provides a sidewall that exhibits significantly reduced sidewall damage compared to conventional methanol-based RIE, as well as minimal residue.

[0026] In the exemplary embodiment, a thickness t remains after the etching transfer of the DARC or BARC layer. This depends on the initial thickness and the composition of the layer. 16and the etching conditions, however, the DARC or BARC layer can be lost during the etching process. 32m completely removed, so that the hard mask surface 15t is exposed. A hard mask is therefore advantageous. 15 selected which has a high etch rate selectivity for the underlying MTJ layers, so that after the etching process 32m A considerable thickness of the hard mask remains.

[0027] An optional volatilization step 34v , who in Fig. The tool shown in figure 6 can be used to remove any residue that has accumulated on the side wall. 20 at the end of the etching step 32m have accumulated. The volatilization preferably comprises IBE or plasma sputtering with an inert gas, wherein Ar +-ions or ions of Kr, He or Ne, generated with a high-frequency or direct current power, orthogonally or at a specific penetration angle onto a top surface 10t are directed towards the substrate. The wafer (not shown) on which the MTJ nanocolumn is placed. 1 The material being formed is typically rotated during IBE or remains static for plasma sputter etching during the step. 34v In another embodiment, volatilization is a thermal treatment in an inert or oxidizing atmosphere at a temperature of 50 °C to 450 °C.

[0028] According to a second embodiment, which is described in the Fig. 4a-4b, Fig. 5 and Fig. As shown in Figure 6, the present disclosure comprises a process sequence in which the chemical component in the combined etching of the first embodiment is separated from the physical etching component. One or two optional steps involving a chemical treatment ( Fig. 5) and volatilization ( Fig. 6) can be included to form an MTJ nanocolumn 1a to be achieved with minimal residue and significantly reduced sidewall damage compared to conventional methanol-based RIE in industry-standard practice.

[0029] In Fig. 4a will be an IBE 32i used with inert gas to create MTJ layers 11-14 to remove those not passed through the DARC or BARC layer 16 and through the hard mask 15 are protected. As a result, the MTJ nanocolumn is protected. 1a produced, but exhibits significant non-volatile residues 19 on the side wall 20and on the substrate surface 10t on, which surround the MTJ nanopillar. After that, as in Fig. 5 shows a chemical treatment 33 This is carried out to convert the non-volatile residues into volatile residues. 19xto convert the residue on the side wall. Without restrictions from the theory, it is assumed that the chemical treatment converts metal residues such as Ta into their oxide, the oxide being more volatile than the metal residue. The chemical treatment involves one or more of methanol, ethanol, and ammonia, and a temperature between room temperature and 150 °C. The chemical treatment takes place in a chamber within an Anelva mainframe or similar device, or it can be performed in a standalone process tool outside of a mainframe, although the latter option slows the throughput. In some cases, the one or more chemicals are introduced into the chamber along with the substrate. Furthermore, oxygen can be added to the chemical treatment chamber to increase the oxidation rate of the non-volatile residue. 19to increase. Alternatively, a plasma can be generated from one or more chemicals under RIE conditions in an etching chamber. Preferably, the applied RF power is so low that no significant damage occurs to the MTJ sidewall. 20 occur. For example, the RF power can be kept between 100 watts and 800 watts.

[0030] With reference to Fig. 6. The volatilization step 34v to be carried out to remove the volatile residues 19x to remove the residues, and includes IBE or plasma sputter etching with Ar or another inert gas flow, or thermal treatment in an etching chamber at a temperature between 50 °C and 450 °C for a duration of up to 5 minutes. In each of the processes involving IBE, plasma sputter etching, or heat treatment, an oxygen flow can be added to the inert gas flow during the volatilization step. The step 34vmay have a preset endpoint time of the duration based on a separate experimental study that determined an evaporation time for a typical thickness of residues 19x determined. If Ar plasma sputtering is selected, the RF power is preferably kept at 100 watts or below to avoid damage to MTJ sidewalls. Preferably, an encapsulation process immediately follows the plasma sputtering in the same process chamber, as described below.

[0031] The present disclosure also includes an embodiment wherein the steps 33 and 34v simultaneously after the etching process 32iThe process can be carried out in various ways. In particular, one or more mixtures of methanol, ethanol, NH3, and CO, along with an inert gas flow, can be introduced into an etching chamber. Thermal treatment can be performed at a temperature between 50 °C and 450 °C while the gas mixture is in the etching chamber. In some embodiments, a plasma is generated using RF power while the gas mixture is in the chamber at a temperature in the range of 50 °C to 150 °C. Alternatively, plasma sputtering etching can be performed at a temperature near room temperature.

[0032] In a third embodiment, which corresponds to a modification of the process sequence of the second embodiment, a reactive ion etching 32r as in Fig. 4b shown, which includes one or more chemicals including methanol, ethanol, NH3 and CO, is applied either before or after step 32i the Fig. 4a inserted. In this case, a chemical treatment is required. 33 possibly not necessary, since the step 32r non-volatile residues 19 can oxidize, which was caused by a previous step 32i were produced to remove volatile residues 19x on the side wall 20 to obtain. The volatilization step 34v can be done after the steps 32i and 32r to be applied to remove all volatile residues that may be on the sidewall 20 the MTJ nanopillar 1a remain.

[0033] In a fourth embodiment, the process sequence in the second embodiment is modified so that the etching step 32i through the etching step 32r is replaced. The step 33 Chemical treatment is usually not necessary, as step 32rtends to serve the same purpose, namely the oxidation of non-volatile residues that are on the side wall 20 to be formed. An optional volatilization step 34v can the step 32r Follow these steps to remove any volatile residues that may be present on the side wall. 20 the MTJ nanopillar 1a form.

[0034] With reference to Fig. 7 becomes an encapsulation layer 25 , which includes a dielectric material, deposited over the MTJ arrangement after clean sidewalls 20formed according to one of the preceding embodiments. Preferably, the encapsulation layer has a thickness of 5-250 nm and is one or more of SiN, SiO2, MgO, Al2O3, AlN, BN or the like, which is deposited immediately after volatilization (in-situ) by physical vapor deposition (PVD), chemical vapor deposition (CVD), ion beam deposition (IBD) or atomic layer deposition (ALD) without interrupting the vacuum in the process chamber previously used for plasma sputtering in the step 34v was used. In embodiments where heat treatment is performed in a separate tool for the step 34v When using this method, the wafer should be exposed to the atmosphere and placed in the encapsulation process chamber.

[0035] Afterwards, a chemical-mechanical polishing process (CMP process) is carried out to create a top surface. 25tto form on the encapsulation layer, which is coplanar with the top surface 15t on the hard mask 15 In some embodiments, the CMP process removes all DARC or BARC layers. 16 , which after the etching transfer steps 32m , 32i or 32r remain in the embodiments described above.

[0036] With reference to Fig. Figure 8 shows the multiple island shapes of the MTJ nanopillars formed by an etching process or process sequence of the present disclosure, after removal of overlying layers in the storage structure, in a top view. The MTJ nanopillar 1a is in the same series as the MTJ nanopillar 1b depicted, and the MTJ nanopillars 1c , 1d are located in the same column as the respective MTJ nanopillar 1aand ib. As discussed above, the MTJ nanopillars are depicted as circular, but can have elliptical shapes in other embodiments. Generally, millions of MTJ nanopillars are formed in an array, but only four are illustrated here for simplicity.

[0037] Subsequently, an upper electrode layer comprising multiple parallel conductive lines (not shown) is deposited onto the MTJ nanopillars and the encapsulation layer using a conventional method. 25 formed, as is standard practice in the industry. A first upper electrode lead can form a top. 15t the MTJ nanopillars 1a , 1c contact, while a second upper electrode line touches the top 15t the MTJ nanopillars 1b , 1dContacts. Conductive traces in the upper electrode layer are preferably formed along the z-axis direction, which runs orthogonally to the conductive traces along the x-axis direction in the lower electrode layer. Therefore, the lower electrode line can 10 one underside of the two MTJ nanopillars 1a and 1b contact, while a second lower electrode line 10-1 the lower surfaces of the MTJ nanopillars 1c and 1d contacted.

[0038] With reference to Fig. Figure 9 provides a flowchart for the process flow of the first embodiment, in which a structure with a CD in the MTJ hard mask layer is produced by an etching process. 100 is formed, which takes the step 30 or both of the steps described above 30 , 31 includes. In step 110 A combined physicochemical etching process will be used. 32mused to transfer the structure through the remaining MTJ layers, thus forming multiple MTJ nanopillars. An optional volatilization step. 114 , which includes IBE, plasma etching or thermal treatment, is performed after the step 110 carried out. Finally, in step 120 An encapsulation layer was formed around the MTJ nanopillars to electrically isolate the MTJ nanopillars from each other.

[0039] In Fig. Figure 10 shows a flowchart for the process flow of the second embodiment of the present disclosure, wherein a physical (IBE) etching 32i in step 111 , a chemical treatment 33 in step 113 , a volatilization process 34v in step 114 and then an encapsulation with step 120 on the step 100 consequences.

[0040] Fig. Figure 11 shows a flowchart for the process flow of the third embodiment of the present disclosure, wherein a chemical (RIE) etching 32r in step 112 , an IBE step 111 , an optional volatilization step 114 and finally an encapsulation in step 120 on the step 100 follow. In the alternative embodiment described above, the IBE step can 111 the RIE step 112 preceded by the optional step 114 and the step 120 .

[0041] We have demonstrated the advantages of the combined physicochemical etching process of the present disclosure with results from an experiment in which an array of MTJ nanopillars with different diameters (w1 in Fig. 8) and reference MTJ nanopillars were fabricated in the same range of apparatus sizes. All process steps described below involved etching a DARC layer or hard mask stack under conditions comprising CF4 RIE, 500 watts (top) / 52 W (bottom), 50 sccm of CF4 only, and 4.5 mT pressure, followed by a passivation step involving the introduction of oxygen into the etch chamber. Each MTJ layer stack in all apparatuses contained an MgO tunnel barrier layer between a CoFeB-free layer and a CoFeB reference layer, a Ta hard mask, and a TaN / NiCr seed layer.

[0042] Following hard mask etching, reference MTJ nanopillars were fabricated using a conventional method employing an Ar IBE comprising a 450 mm IBS (ion beam source) and 800 W IBS RF power, 200 V / 950 V Gi / G2 voltage, 400 mA Gi current, 60 rpm rotation, and penetration angles of 40° and 80° without a subsequent volatilization process. The wafers were exposed to air between MTJ etching and encapsulation.

[0043] According to a process described in the first embodiment, an MTJ stack of layers was 11-14 ( Fig. 3) with an overlying structured hard mask layer 15The surface was etched with a RIE step comprising a 50% Ar / 50% methanol mixture and an RF power of 1500 watts (top) / 1100 watts (bottom) for 60 seconds. The Ar flux rate was 7.5 standard cubic centimeters per minute (sccm) and the CH3OH flux rate was 7.5 sccm. This was followed by a volatilization step comprising Ar plasma sputtering at an RF power of 75 watts and an Ar flux rate of 60 sccm for 132 seconds at room temperature.

[0044] Fig. Figure 12 shows a diagram of the magnetoresistive ratio (DRR) plotted against the MTJ size (measured at 125 °C) for the reference MTJ nanopillar, which was structured using the IBE step. The size of the MTJ nanopillar ranges from approximately 30 nm to 300 nm. The outlined area 60shows a significant population of low tails corresponding to devices where redeposited metal residues bridge the MgO tunnel barrier layer and cause short circuits.

[0045] In Fig. Figure 13 shows the results for MTJ nanopillars etched using the 50 / 50 Ar / CH3OH etching method and then subjected to a volatilization step using Ar plasma sputtering. The number of devices with low DRR, and thus a higher overall DRR, especially for MTJ nanopillar sizes between 30 and 100 nm, is significantly reduced. The DRR is also more uniform for each MTJ size. Fig. 13 compared to Fig.12, which indicates a more controllable process suitable for a production environment. Furthermore, the embodiments of the etching process disclosed herein can be easily implemented in existing production lines, as no new tools or materials are required.

[0046] Although this disclosure has been shown and described specifically with reference to the preferred embodiment, the person skilled in the art will understand that various changes in form and details can be made without departing from the spirit and scope of this disclosure. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] US 8722543

[0001]

Claims

[1] Method for etching a magnetic tunnel junction layer stack (MTJ layer stack) comprising: (a) Providing an MTJ layer stack on a first electrode, the MTJ layer stack comprising a top hard mask layer and a first layer stack with a reference layer, a free layer and a tunnel barrier layer between the reference layer and the free layer; (b) Forming a structure in the hard mask layer by a first etching step which is ion beam etching (IBE) with an inert gas or reactive ion etching (RIE) with a fluorocarbon or carbon chloride, wherein the structure has a sidewall extending from a hard mask top to a top of the first layer stack; and (c) Forming a structure in the first layer stack with a sidewall forming a continuous surface with the sidewall in the hard mask layer and extending to a top surface of the first electrode, wherein the structure in the first layer stack is produced by a second etching step combining a physical component in the form of noble gas ions or plasma and a chemical component in the form of ions or plasma of one or more chemicals selected from methanol, ethanol, ammonia, N2O, H2O2, H2O and carbon monoxide. [2] Method according to claim 1, further comprising a volatilization step for removing volatile residues on the first stack side wall after the second etching step. [3] Method according to claim 2, wherein the volatilization step is an IBE or a plasma sputtering etching comprising an inert gas and an RF power or a DC power. [4] Method according to claim 3, wherein the IBE or plasma sputtering etching during the volatilization step generates inert gas ions or a plasma which are directed with a specific penetration angle between 0° and 90° for the IBE, or orthogonally with respect to a top surface of the first electrode for the plasma sputtering etching. [5] Method according to claim 2, wherein the volatilization step is a thermal treatment comprising a temperature between about 50 °C and 450 °C. [6] The method of claim 1, wherein the one chemical or the several chemicals in the second etching step further comprise oxygen. [7] Method according to claim 1, wherein a direction of ions and plasma in the second etching step is substantially orthogonal to the top of the first electrode. [8] Method according to claim 1, wherein the ions and the plasma are generated in the second etching step with an RF power in the range of 600 to 3000 watts. [9] Method according to claim 5, wherein the volatilization step further comprises a flow rate of an inert gas, or a flow rate of oxygen and an inert gas. [10] Method for etching a magnetic tunnel junction layer stack (MTJ layer stack) comprising: (a) Providing an MTJ layer stack on a first electrode, the MTJ layer stack comprising a top hard mask layer and a first layer stack with a reference layer, a free layer and a tunnel barrier layer between the reference layer and the free layer; (b) Forming a structure in the hard mask layer by a first etching step which is ion beam etching (IBE) with an inert gas or reactive ion etching (RIE) with a fluorocarbon or carbon chloride, wherein the structure has a sidewall extending from a hard mask top to a top of the first layer stack; and (c) Forming a structure in the first layer stack with a sidewall that forms a continuous surface with the sidewall in the hard mask layer and extends to a top surface of the first electrode, wherein the structure in the first layer stack is produced by a process sequence comprising: (1) a first step which is an IBE with an inert gas; and (2) a second step, which is a chemical treatment to convert a non-volatile residue formed on the continuous surface during the first step into a volatile residue. [11] The method of claim 10, further comprising carrying out a volatilization step to remove the volatile residue on the continuous surface after the chemical treatment. [12] Method according to claim 11, wherein the volatilization step comprises a second IBE step or a second plasma sputtering etching step with an inert gas and an RF power of less than 100 watts. [13] Method according to claim 11, wherein the volatilization step comprises one or both of a heat treatment at a temperature between about 50 °C and 450 °C and an introduction of oxygen into the chamber. [14] Method according to claim 10, wherein the chemical treatment comprises one or more of methanol, ethanol, NH3 and CO. [15] Method according to claim 14, wherein the chemical treatment further comprises oxygen. [16] Method according to claim 10, wherein the second step further comprises the addition of an inert gas and a temperature in the range of 25 °C to 150 °C to remove the volatile residue. [17] Method according to claim 14, wherein the chemical treatment further comprises RIE conditions with an RF power of about 100 to 800 watts generating a plasma. [18] Method for etching a magnetic tunnel junction layer stack (MTJ layer stack) comprising: (a) Providing an MTJ layer stack on a first electrode, the MTJ layer stack comprising a top hard mask layer and a first layer stack with a reference layer, a free layer and a tunnel barrier layer between the reference layer and the free layer; (b) Forming a structure in the hard mask layer by a first etching step which is ion beam etching (IBE) with an inert gas or reactive ion etching (RIE) with a fluorocarbon or carbon chloride, wherein the structure has a sidewall extending from a hard mask top to a top of the first layer stack; and (c) Forming a structure in the first layer stack with a sidewall forming a continuous surface with the sidewall in the hard mask layer and extending to an upper surface of the first electrode, wherein the structure in the first layer stack is produced by a second etching step comprising RIE conditions with one or more chemicals selected from methanol, ethanol, N2O, H2O2, H2O, ammonia and carbon monoxide. [19] Method according to claim 18, further comprising carrying out a volatilization step to remove a volatile residue from the first stack side wall after the second etching step, wherein the volatilization step comprises IBE or plasma sputter etching with an inert gas. [20] Method according to claim 18, further comprising carrying out a volatilization step after the second etching step, wherein the volatilization step comprises a temperature in the range of 50 °C to 450 °C. [21] Method according to claim 19, wherein the IBE or plasma sputtering etching further comprises the introduction of an oxygen flow with the inert gas. [22] Method according to claim 20, wherein the volatilization step further comprises the introduction of an oxygen flow.