A single milling angle, φ1, is typically selected for defining the conventional magnetoresistive junction 10. This milling angle is typically at least five degrees and not more than thirty degrees from normal to the surface of the transducer. The single-angle milling proceeds until the stack has been completely milled through. Thus, the conventional magnetoresistive junction 10 is substantially defined. Because this single-angle ion milling often leads to redeposition of removed material on the sides of the conventional junction 10 and mask, a second, cleanup ion mill may be performed. This second ion mill is typically short in duration and performed at a high angle milling angle, φ2. For example, the angle is typically greater than sixty degrees from normal to the surface of the read transducer.
Although the conventional ion milling may define the conventional magnetoresistive junction 10, there are drawbacks. Ion milling may cause damage to the layers in a stack, particularly to oxide layers. Thus, the first, single-angle ion mill may damage the barrier layer 16 when the conventional magnetoresistive junction 10 is defined. This damage to the barrier layer 16 may adversely affect performance of the magnetoresistive junction 10. In addition, if the redeposition is not cleaned by the second ion milling, then metallic redeposition across the barrier layer 16 may result in shorting of the magnetoresistive junction 10. However, if the redeposition is cleaned, then the additional ion mill may further damage the barrier layer 16.
In addition, the damage due to ion milling may vary based on junction angle, θ, of the magnetoresistive junction 10 as well as on the milling angle, φ. The ion milling damage to the barrier layer 16 of the conventional magnetoresistive junction 10 occurs when the junction width is close to its final value. This is because portions of the stack damaged far from the final width of the conventional magnetoresistive junction 10 are removed during ion milling. As a result, ion milling damage is generally smaller for a shallow magnetoresistive junction (small junction angle θ and large milling angle φ) than for a steep magnetoresistive junction (large junction angle θ and small milling angle φ). This is because the shallow junction 10 is typically milled using a high milling angle φ1. As a result, the junction width for a shallow magnetoresistive junction 10 reaches its final value only when the single-angle milling comes towards its end. In contrast, for a larger junction angle θ formed using a small milling angle φ1, the width quickly gets close to its final value. As a result, the barrier layer 16 is exposed to more milling during the single-angle mill and experiences greater damage. Thus, a conventional magnetoresistive junction that has a steep (large) junction angle and/or which is formed using a small milling angle is more likely to be damaged during ion milling that defines the junction.
The conventional ion mill process may also create an undesirable junction profile. The single-angle ion mill or the single-angle ion mill in combination with the second ion mill may result in a kink 19, or step, at the barrier layer 16. This profile is due to the redeposition during the single-angle ion mill and different milling rates of the stack layers. For example, the barrier layer 16 typically mills at a different rate than the pinned layer 14 or free layer 18. Consequently, especially for a shallow junction angle, the kink 19 may occur. This junction profile with a kink 19 at the barrier layer 16 is undesirable because it adversely affects biasing of the magnetoresistive junction 10 by the hard bias structure (not shown). Consequently, performance of the read transducer 10 may be adversely affected.
Further, the trend in magnetic recording is to higher densities and, therefore, smaller junction widths. For example, current ultra-high density magnetic recording of approximately five hundred GB/in2 or more utilizes a TMR junction 10 having a width of not more than fifty nanometers. The junction width is desired not only to be small, but to have limited variations in order to maintain performance. Using the conventional single-angle ion milling, the junction width is primarily determined by the width of the mask used during ion milling. This is generally true whether or not the second ion mill is performed. At higher densities, the photolithography utilized to repeatably obtain a mask having a small width with limited variations may be difficult to achieve. Consequently, fabrication of the conventional magnetoresistive junction 10 may be more problematic.
There are conventional mechanisms for accounting for ion mill induced damage. Damage caused by the single angle ion mill that defines the junction and the second, cleanup ion mill may be repaired by an oxidation. However, such an oxidation may result in a relatively thick oxidation layer on the sides of the conventional magnetoresistive junction 10. Consequently, biasing of the magnetoresistive junction using a hard bias layer (not shown in
Accordingly, what is needed is a system and method for providing an improved magnetoresistive junction.
A method and system define a magnetoresistive junction in a magnetic recording transducer. The method and system include performing a first mill at a first angle from a normal to the surface of the magnetic recording transducer. A second mill is performed at a second angle from the normal to the surface. The second angle is larger than the first angle. A third mill is performed at a third angle from the normal to the surface. The third angle is not larger than the first angle.
A first mill is performed at a first angle from normal to the surface of the read transducer, via step 102. Thus, for an ion mill performed in step 102, the ions are incident on the magnetoresistive stack at the first angle from the normal. In one embodiment, the first angle is at least twelve degrees and not more than thirty degrees from normal. In one such embodiment, the first angle is at least seventeen degrees and not more than twenty-five degrees from normal. In one embodiment, the first mill is terminated after at least a portion of the junction including the barrier layer is defined. Thus, the first mill exposes at least the free layer and the barrier layer in a TMR junction in which the pinned layer is closer to the underlying substrate. In one such embodiment, the first ion mill is terminated after the layer immediately below the barrier layer is defined. Stated differently, the first mill would be terminated before another portion of the magnetoresistive junction including the layer immediately below the barrier layer is completely defined. In one embodiment, therefore, the first mill may be terminated before the pinned layer is milled through and this portion of the junction completely defined.
After termination of the first mill, a second mill is performed at a second angle from normal to the surface of the read transducer, via step 104. For an ion mill performed in step 104, the ions are incident on the magnetoresistive stack at the second angle from the normal. This second angle is greater than the first angle. In one embodiment the second angle is at least sixty degrees and not more than eighty degrees. In one such embodiment, the second angle is at least seventy degrees from normal. In one embodiment, the second mill is terminated before the junction is completely defined. Thus, in such an embodiment, the second mill is terminated before the pinning layer is completely milled through.
A third mill is performed at a third angle from normal to the surface of the read transducer after termination of the second mill, via step 106. Thus, for an ion mill performed in step 106, the ions are incident on the magnetoresistive stack at the third angle from the normal. The third angle is not larger than the first angle. In one embodiment, the third angle is smaller than the first angle. In one embodiment, the third angle is not more than twelve degrees. In one such embodiment, the third angle is at least three degrees and not more than nine degrees from normal. In one embodiment, the third mill is terminated after the magnetoresistive junction is completely defined.
Using the method 100, the magnetoresistive junction may be defined. Moreover, the magnetoresistive junction, particularly the barrier layer, may exhibit less damage. The first mill may be performed at a relatively large angle. The second mill is performed at an even larger angle. As described above, a larger angle from normal results in less damage to the junction. Using the method 100, therefore, less damage may be done to the junction while a significant portion of the junction is being defined. For example, in one embodiment, at least the free and barrier layers are defined substantially defined in the first and second mills. Thus, these layers may exhibit less damage due to ion milling. Further, the second mill may be performed at a sufficiently high angle to remove redeposition that has built up during the first mill. Thus, less damage and less redeposition may be result in the final device. Because less damage may be done during definition of the magnetoresistive junction, oxidation steps meant to repair such damage may be skipped or reduced in strength. For example, a natural oxidation instead of a plasma oxidation may be sufficient. Consequently, processing may be simplified and thick oxide layers at the sides of the junction may be reduced or avoided. Furthermore, the third mill may be performed at a lower angle from normal to the surface. Although this third mill may be more likely to damage the magnetoresistive junction, it may be performed for a relatively short time. This is because the first two mills have already defined a significant portion of the junction. The third mill allows the junction angle and the width for the magnetoresistive junction to be tailored substantially as desired. In particular, a steeper junction may be achieved. Further, the width of the magnetoresistive junction may be adjusted in the second mill without reducing the size of the mask used in defining the junction. As a result, photolithography parameters may be relaxed. Fabrication may, therefore, be simplified.
A first ion mill is performed at a first angle, Φ1 of at least twelve and not more than thirty degrees from normal to the surface of the read transducer, via step 152. In one embodiment, the first angle is at least seventeen degrees and not more than twenty-five degrees from normal. Also in step 152, the first mill is terminated after at least a portion of the junction including the barrier layer 224 is defined but before the pinned layer 222 has been completely milled through.
After termination of the first ion mill, a second ion mill is performed at a second angle of at least sixty degrees from normal to the surface of the magnetoresistive stack 210′, via step 154. In one such embodiment, the second angle is at least seventy degrees and not more than eighty degrees from normal. Also in step 154, the second mill is terminated before the pinning layer 222 is completely milled through.
After termination of the second mill, a third mill is performed at a third angle of not more than nine degrees from normal to the surface of the magnetoresistive stack 210″, via step 156. The third mill continues in step 156 until the magnetoresistive junction is completely defined.
Using the method 150, the magnetoresistive junction 210′″ may be defined. Moreover, the magnetoresistive junction 210′″, particularly the barrier layer 226″, may exhibit less damage. Because less damage may be done during definition of the magnetoresistive junction, oxidation steps meant to repair such damage may be skipped or reduced in strength. For example, a natural oxidation instead of a plasma oxidation may be sufficient. Consequently, processing may be simplified and thick oxide layers at the sides of the junction may be reduced or avoided. In addition, redeposition 230 has been removed. Furthermore, a larger junction angle may be achieved and the width of the magnetoresistive junction 210′″ adjusted. As a result, photolithography parameters may be relaxed. Fabrication may, therefore, be simplified. In addition, as can be seen in
In addition, an oxide layer 272 is shown. Because the method 100 or 150 is used, the oxide layer 272 may be thin. In particular, the oxide layer 272 is not more than one nanometer thick at the free layer (not shown in
Because the method 100 or 150 is used, the read sensor 260 may exhibit less damage. Further, little or no redeposition may reside on the read sensor 260. Thus, shorting of the read sensor 260 may be less likely. The junction angle and track width of the read sensor 260 may also be better controlled. As a result, photolithography parameters may be relaxed. Fabrication may, therefore, be simplified. Further, because less damage may be done during definition of the read sensor 260, oxidation steps meant to repair such damage may be skipped or reduced in strength. For example, a natural oxidation instead of a plasma oxidation may be sufficient. For example, the oxide layer 272 may have a thickness of less than one nanometer. Consequently, the read sensor 272 may be better coupled with the hard bias 270. Performance of the magnetic head 250 may, therefore, be improved.
This application is a divisional of U.S. patent application Ser. No. 12/147,450, filed on Jun. 26, 2008, which is hereby incorporated by reference in its entirety.
Number | Name | Date | Kind |
---|---|---|---|
6538858 | Hasegawa et al. | Mar 2003 | B1 |
6643107 | Hasegawa et al. | Nov 2003 | B1 |
6663987 | Kakihara et al. | Dec 2003 | B2 |
6700759 | Knapp et al. | Mar 2004 | B1 |
6713237 | Seigler et al. | Mar 2004 | B2 |
6795280 | Song et al. | Sep 2004 | B1 |
6847508 | Tanaka et al. | Jan 2005 | B2 |
7002782 | Hasegawa et al. | Feb 2006 | B2 |
7042684 | Horng et al. | May 2006 | B2 |
7043823 | Childress et al. | May 2006 | B2 |
7081041 | Guthrie et al. | Jul 2006 | B1 |
7194797 | Pinarbasi | Mar 2007 | B2 |
7212384 | Stoev et al. | May 2007 | B1 |
7218485 | Hasegawa et al. | May 2007 | B2 |
7218487 | Hasegawa et al. | May 2007 | B2 |
7268985 | Freitag et al. | Sep 2007 | B2 |
7367110 | Cyrille et al. | May 2008 | B2 |
7639459 | Childress et al. | Dec 2009 | B2 |
8031444 | Chou et al. | Oct 2011 | B2 |
8432645 | Matsuzawa et al. | Apr 2013 | B2 |
20020015268 | Mao et al. | Feb 2002 | A1 |
20030030947 | Ooshima | Feb 2003 | A1 |
20030231437 | Childress et al. | Dec 2003 | A1 |
20040223267 | Childress et al. | Nov 2004 | A1 |
20040264068 | Kanaya et al. | Dec 2004 | A1 |
20050264957 | Li et al. | Dec 2005 | A1 |
20060067009 | Cyrille et al. | Mar 2006 | A1 |
20060094129 | Pinarbasi | May 2006 | A1 |
20060291097 | Honda et al. | Dec 2006 | A1 |
20070056160 | Pinarbasi | Mar 2007 | A1 |
20070081279 | Hong et al. | Apr 2007 | A1 |
20070266549 | Gill et al. | Nov 2007 | A1 |
20090059443 | Tsuchiya et al. | Mar 2009 | A1 |
20090067099 | Tsuchiya et al. | Mar 2009 | A1 |
20090086383 | Hara et al. | Apr 2009 | A1 |
Number | Date | Country | |
---|---|---|---|
20130216863 A1 | Aug 2013 | US |
Number | Date | Country | |
---|---|---|---|
Parent | 12147450 | Jun 2008 | US |
Child | 13848353 | US |