Rotary counter using a magnetic wall conduction path wound in a closed loop

Abstract

The present invention relates to a rotary counter using a domain wall conductive path wound in a closed loop. In particular, to solve the problem of providing a rotary counter with the largest possible magnetic window, the connection area where the inner end and the outer end of the loop part of the loop of the first domain wall conductive path (M1) are placed together is bridged by a second domain wall conductive path (M2) attached to both ends of the spiral domain wall conductive path through one air gap (201), the air gap (201) forms a local interruption in the domain wall conductive path (M1) to be closed, the average width of the air gap (201) is determined to be smaller than the thickness of the domain wall conductive path (M1, M2), and the adjacent domain wall conductive path parts (M2 or M1) are covered by a non-magnetic layer in the air gap area.

Description

Technical Field

[0001] The present invention relates to a rotary counter using a magnetic wall conduction path wound in a closed loop, and is particularly used for a rotary counter for a particularly large rotation speed U (U>100).

Background Art

[0002] A magnetic sensor can detect characteristics of a magnetic field, such as the direction of the magnetic field. An example thereof is an angle sensor. Another type of magnetic sensor can determine the frequency at which the magnetic field rotates. Such a rotary counter can be formed, for example, by a well-known prior art GMR rotary counter or TMR rotary counter (for example: RSM-2800: https: / / www.novotechnik.de / fileadmin / user_upload / pdfs / kataloge_flyer / Flyer_RSM-2800.pdf). The basic principle of such a magnetic rotary counter is based on the use of a magnetic wall moving within a continuous magnetic region that can be regarded as a magnetic conductor. The magnetic wall is made of a soft magnetic metal material. The reading principle utilizes the magnetoresistance effect. For this purpose, another magnetic layer and a non-magnetic layer are required that enable the magnetic wall to be localized at least locally. Such a rotary counter is technically realized as follows. A magnetic wall is generated by an external magnetic field (for example, a rotating permanent magnet) in an electrically and magnetically continuous magnetic conductor that is itself part of a GMR stack or locally forms a TMR stack, and moves within a spirally arranged magnetic conductor. In this case, the length of the magnetic conductor increases in proportion to the number of turns N forming the spiral because the length of each turn increases towards the outside. Assuming the length of the innermost loop is L1 and the distance between each turn is a, the length L i of the i-th turn of the spiral composed of N turns is calculated using the following formula. That is, the length L(i)=L1+8(i-1)·a. Thus, the total length L(N) of the spiral with N turns is the sum of the individual lengths L i and is L(N)=N·L1+4·a·(N 2-N). Therefore, this length L(N) increases in proportion to N or more linearly. As the length of the continuous magnetic conductor increases, the probability that defects that impede the movement of the magnetic wall and render its function unfulfilled exist in the conductor also increases proportionally. As a result, the yield decreases disproportionately as N increases. At the same time, the required chip area becomes increasingly larger as N increases, mainly because the number of contacts required for reading increases in direct proportion to N. Due to this characteristic, using an open spiral that is not closed to actually realize a rotation counter is limited to the case of N ≤ 64. In the technical field, there are many requirements to count rotation speeds far exceeding 64 rotations, which cannot be realized by the basic principle described above.

[0003] As proposed in DE102013018680A1, the counting of the rotation speed can also be realized not by using a single open spiral, but by using a plurality of closed loops CL i (CL = closed loop), and configuring the i-th loop with a spiral having N i turns. The outer ends and inner ends of all the spirals N i are connected to each other to form one closed loop CLi respectively. Furthermore, according to Patent DE102013018680A1, the loops Ni for i = 1...4 are designed such that they are relatively prime to each other. Thus, the rotation counter designed in this way can count from 1 to N1·N2·...N n up to. In the exemplary case of n = 4 loops with Ni = 5, 7, 9, and 13, this means that a rotation speed of up to 4095 rotations can be obtained. In this case, the total length of all the closed loops corresponds to a single spiral with N = 34 according to the above solution. What becomes clear from this is CL iSince the total number of individual loops in each loop is significantly small, the total length of all loops can also be designed to be significantly short. This also means that, in order to realize the rotation counter in this way, the total length of the magnetic conduction path that determines the yield is significantly shortened. At the same time, in this proposal, the number of connection pads required to read out each rotation information can also be reduced. As a result, furthermore, the surface area of this rotation counter can be kept smaller, and thus it can be manufactured at a lower cost. To realize the closed loop CLi, the inner end and the outer end of all loops CL i must be connected to each other. Two solutions are described in the prior art for this. In DE102013018680A1, as shown in FIG. 1, it is proposed to realize the connection by introducing a new functional element, i.e., an intersection in the magnetic conduction path. For this solution, the entire rotation sensor is arranged on one plane. The advantage of this solution is that the intersection can also be manufactured in the same technical step as making the spiral. However, the disadvantage of this proposed solution is that in the intersection region, the width of the magnetic conduction path increases to at least a value of 1.414 times (the diagonal D of both webs of the width w forming the intersection region in FIG. 1), so that in the structure, the width of the magnetic conduction path is from w to D exp in the range. The important characteristic of the sensor, i.e., the magnetic field region in which the sensor can count without error, is directly related to the shape. There is a minimum magnetic field B min that must exist in order to always reliably transport the magnetic wall in the magnetic wall conduction path so that it functions without error. Similarly, the magnetic field B max must not be exceeded. Because otherwise, additional magnetic walls will be generated uncontrollably in the structure. Since the values of B min and B max are indirectly proportional to the width w, at the intersection where the width value increases by 1.42 times, the minimum induction B min and the maximum induction B maxThe value leads to a local decrease. This is evident in Figure 2. Here, it shows how the magnetic window ΔB depends significantly on the strip width w. In Figure 2, for a spiral with a strip width w = 350 nm, the magnetic window ΔB is indicated by the double-dotted line. The magnetic window ΔB at the point where w increases to 525 nm is represented by the vertical double-dashed line. Since there are structures with w = 350 nm and w = 525 nm within the spiral with intersections, as shown in Figure 2, the magnetic window ΔB occurs as the vertical interval between two parallel horizontal lines. From this, a rotational sensor with intersections can only operate within the significantly narrow range of ΔB shown in Figure 2, which is formed by the small value of B max and the large value of B min . B min and B max The difference represents the width of the magnetic window where the rotation counter can be used. In the case shown in Figure 2, when the diagonal D successfully realizes an intersection with a length of 525 nm, the width of the magnetic window ΔB will decrease from 15 mT (for the open spiral with b = 350 nm) to only 5 mT. In that case, the magnetic window will be limited by the B max value when D = w = 525 nm and the B min value when w = 350 nm, and will drop from 15 mT to 5 mT compared to the spiral without intersections. As shown in Figure 1, in the known realized loops, the shape of the intersection (see circle A2 in Figure 1) deviates from the ideal shape (circle A1). Due to the rounding of the intersection, the value of D increases. Since the value D exp is larger than the value D of the ideal intersection, the experimentally achievable magnetic window with a magnitude of 1 - 2 mT is expected to be even smaller than the 5 mT in Figure 2, and thus is completely unsuitable for practical use. This magnetic window should be as large as possible for reasons of application technology in order to determine what interference magnetic field resistance the sensor has. If the magnetic window is narrow, in order to suppress unacceptable interference magnetic fields, although not described in detail here, additional significant technical effort is required, leading to an increase in the cost of the sensor system and an increase in the size of the design shape. At the same time, a narrow magnetic window requires a narrow tolerance with respect to the installation position in the manufacture of the rotary counter, and also requires a narrow tolerance for the magnet 12 that generates the rotating magnetic field, as shown in the possible application examples in FIG. 3, which also leads to an increase in cost.

[0004] Furthermore, a closed-loop structure without intersections was proposed in patent DE102010022611B4. This solution requires that the short-circuit of the spiral, i.e., the connection between the inner end and the outer end of the spiral by the magnetic conductor, be designed so that no intersections occur. This can only be achieved by realizing this connection in a plane E2 above or below the plane E1 in which the open spiral is located (see FIGS. 4 and 5). In the ideal case where the width of the magnetic conductor is constant everywhere, the embodiment of the rotary counter with this closed loop will have a magnetic window with the same width as the spiral, and therefore will be significantly wider than the magnetic window that is in principle possible in the example of the intersection described above. However, the technical implementation of the last-mentioned solution is not known to date. The reason is that this solution assumes that the still-open spiral fabricated in plane 1 is provided with an additional magnetic conductor M2 that extends from plane 1 through plane 2 and back to plane 1 as shown in the top view of FIG. 4 and the cross-sectional view of FIG. 5, and this connects the two ends of the open spiral in plane 1 at designated locations K1 and K2. The thickness of the layer structure (magnetic wall conduction path) responsible for the movement of the magnetic wall is t. As shown in the left cross-sectional view of FIG. 5, the connection region indicated by M2 is partially located above the plane E1 in which the intersecting spiral is located, or, in a similarly possible solution, below the plane E1, and therefore does not contact the spiral. Only at the locations K1 and K2 where M2 and M1 are in contact are both structures on the common plane E1. The problem with this solution is, firstly, that the short circuit should ideally be designed such that the cross-section of the magnetic conductor M2 forming the short circuit is the same as that of the spiral magnetic conductor M1. This should be relatively easy to achieve in terms of manufacturing technology. What is particularly problematic with this solution is the location where the magnetic conductor M1 and the magnetic conductor M2 are in contact. This has to be done with substantially no offset either in the lateral direction or in the vertical direction. Because, as has been discovered, even a 15-nm displacement in the lateral direction with respect to the support strongly promotes the pinning of the moving magnetic wall, the value of the lower magnetic window B shown in FIG. 2 min will increase. Therefore, the lateral offset has to be less than 1 / 20 of the width of the magnetic conductor in order not to significantly deteriorate, for example, the lower magnetic window, which is about 350 nm for conventionally used sensors. What is further problematic is the requirement that, in order to obtain a continuous magnetic conductor having the same important characteristics for the sensor functions of magnetic wall generation and magnetic wall movement through the magnetic conductor over the entire length, the thickness at the location where both conductors M2 and M1 are in direct contact should also have a minimal difference with respect to the thickness of the conduction path portions M1 and M2 connecting to each other. However, this requirement rules out a possible configuration of providing an overlap at the connection location as shown in the enlarged circle A6 in FIG. 5. FIGS. 1 to 5 illustrate a solution according to the known prior art and the resulting problems.

Prior Art Documents

Non-Patent Documents

[0005]

Non-Patent Document 1

Patent Documents

[0006]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0007] The problem of the present invention is that it does not have the problems of the solutions described in the prior art, or has them only to a much smaller extent, is particularly easy to handle technically, and has a loop-shaped, substantially planar and closed magnetic wall conduction path with as large a magnetic window as possible. The object is to provide a rotary counter using such a magnetic wall conduction path.

Means for Solving the Problems

[0008] The present invention is solved by the features of claim 1, wherein a connection region where the inner end and the outer end of the loop portion of the loop of the first magnetic wall conduction path M1 are placed together is bridged by a second magnetic wall conduction path M2 attached to both ends of the spiral magnetic wall conduction path via one gap each. The gap 201 forms a local interruption in the magnetic wall conduction path M1 to be closed. In this gap 201, when the magnetic wall DW moves from the first magnetic wall conduction path portion (M2 or M1), a stray magnetic field is generated in the first magnetic wall conduction path portion (M2 or M1), and the width is such that it causes nucleation of the magnetic wall DW in the magnetic wall conduction path portion (M2 or M1) following the gap 201 in the moving direction. The average width of the gap 201 is determined to be smaller than the thickness t of the magnetic wall conduction path M1, and the adjacent magnetic wall conduction path portions (M2 or M1) are covered by the non-magnetic layer S1 in the gap region. Advantageous embodiments are the subject matter of the dependent claims.

[0009] The essence of the present invention is to form the magnetic conduction path where the magnetic wall moves, not as a continuous and unbroken conduction path as is common in the prior art, but as shown in FIGS. 6 to 14. In particular, at the transition points from the magnetic wall conduction path M1 to M2 respectively, a narrow gap 201 is provided according to the present invention, and this gap 201 can be filled with a non-magnetic material as illustrated in FIGS. 9 to 11 for example. It has been discovered that the interruption of the magnetic conductor results in a surprising advantage leading to an improvement in the magnetic characteristics of the rotary counter compared to the conventionally known solutions. As revealed by the implemented micromagnetic simulation calculations, the non-excessive gap 201 within the magnetic conductor does not cause any impairment to the functionality of the sensor contrary to expectations. FIG. 7 shows how the magnetic wall DW moves across the gap according to the present invention. The series of images in FIG. 7 schematically show in a top view the magnetization distribution in M1 and M2 when the magnetic wall passes through the gap 201. The topmost figure shows the initial configuration. The magnetization in both regions M1 and M2 is directed to the left. The narrow gap 201 between M1 and M2 is covered by the stray magnetic field proceeding from the right conductor to the left conductor. With a magnetic wall DW existing in the left conductor M1, when a substantially right-directed uniform magnetic field (generated by the external magnet 12 shown in FIG. 3) represented by the thick arrow 13 acts on this arrangement, the magnetic wall DW moves to the right within the left conductor M1. All the partial views marked (1) in FIG. 7 represent the movement of the magnetic wall within the conductor M1. Under the influence of a sufficiently large magnetic field, the magnetic wall moves towards the gap 201. There, the magnetic wall disappears on the right side of the left conductor ((2) shown), which brings about a change in the stray magnetic field in the gap. Together with the external magnetic field generated by the external magnet 12 shown in FIG. 3 and indicated by the thick arrow in FIG. 7, the changed stray magnetic field causes the nucleation of the magnetic wall DW at the left end of the right conductor M2 ((3) shown). This magnetic wall moves further to the right within the right conductor under the influence of the magnetic field ((4) shown). By the disappearance of the magnetic wall on one side and the formation of a nucleus on the other side, the fixation of the magnetic wall that would be observed at the actual transition point between M1 and M2 without a gap in the continuous magnetic conductor according to the prior art described above is prevented.

[0010] The following examples are used to explain the present invention in detail.

Brief Description of the Drawings

[0011]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

DETAILED DESCRIPTION OF THE INVENTION

[0012] Figures 1 to 5 represent known or conceivable prior art, which has already been fully described in conjunction with the relevant problems at the beginning.

[0013] The detailed description of the present invention begins with Figure 6. Figure 6 shows a partial cross-sectional view of a magnetic wall conduction path having a gap 201 according to the present invention between magnetic wall conduction path portions M1 and M2, respectively. On the right side, a transition portion having a gap between M2 and M1 is shown enlarged in partial view A5.

[0014] How the magnetic wall DW progresses across the gap 201 within both magnetic conductor portions is shown in detail in Figure 7. The position of the magnetic wall DW is schematically indicated by a black ellipse, and the magnetization directions within the magnetic wall conduction paths M1 and M2 are indicated by black arrows. The thick arrow represents the magnetic field by which the permanent magnet 12 (see Figure 3) acts on the structure. How the magnetic wall transfer is carried out in detail across the gap 201 according to the present invention has already been described in detail above.

[0015] The right side of Figure 8 shows a typical structure known to be used in a GMR or TMR stack. Layer 103 is an antiferromagnetic layer such as PtMn, and the layer 102 above it is an artificial antiferomagnetic layer AAF (artificial antiferomagnetic layer). This is a layer system known as an artificial antiferromagnet and typically has a structure of CoFe / 0.8nmRu / CoFe. The M1 layer is the layer in which magnetic domains move within the aforementioned loop. This is usually soft magnetic and consists of Ni, Fe and / or Co alloys, such as Ni 81 Fe 19 , NiFeCo, NiFeB, CoFe, CoFeB or combinations of these materials. According to the present invention, there is a condition that the layer M1 is not directly connected to the layer or stack forming M2.

[0016] Figure 9 shows an example where layer M1 and layer M2 are separated by a non-magnetic separation layer S1 indicated by hatching. According to the present invention, this layer is made of a non-magnetic material and can be either conductive, semi-conductive, or insulating. This allows for great variability in technical implementation. Figure 9 shows contacts where the ends of layer M1 and layer M2 are formed obliquely but parallel to each other and geometrically separated by the separation layer S1.

[0017] As shown in Figure 10, the lateral distance between M1 and M2 does not have to be constant within the scope of the present invention and can also vary as in the case of the V-shaped gap shown there. In this technical solution, both regions are cut obliquely in the vertical direction, a minimum distance is provided by the separation layer S1, and then it is covered with the separation layer S2 to completely fill the gap. Also in this solution, both separation layers S1 and S2 can be formed of a non-magnetic material. Non-magnetic means that it must not have a permanent magnetic moment and thus must not be ferromagnetic or ferrimagnetic. That is, these materials must be diamagnetic, paramagnetic, or antiferromagnetic.

[0018] However, the gap between M1 and M2 can also be arranged obliquely in the plane, as illustrated in the top view of the magnetic wall conduction path in Figure 11. Combinations of the various embodiments of the gap shapes described above are also within the scope of the present invention.

[0019] A great advantage for technical implementation is that the solution according to the present invention does not require the cross-sections of M1 and M2 to exactly transition into each other and can be slightly displaced both laterally and vertically, as illustrated in Figures 12 and 13. Figure 12 illustrates a cross-sectional view of a contact where a vertical offset vs of size is allowed. As long as the vertical offset vs is less than 25% of the thickness t of the domain wall path M1, i.e. less than 10 nm for a typical thickness of the soft magnetic layer of 40 nm, the domain wall can jump this gap, as shown diagrammatically in Figure 7. The same is true for the lateral offset hs of both structures M1 and M2, as shown in Figure 13. As long as the lateral offset hs is less than 25% of the width of the domain wall path, i.e. less than 90 nm for a typical system (360 nm), the domain wall can jump the gap between M1 and M2 as described above.

[0020] The proposed solution offers yet another degree of technical freedom. Figure 14 illustrates a solution in which the values ​​of the saturation magnetization of both regions M1 and M2 are different. That is, what is crucial for the function of the domain wall moving through the air gap according to the invention is the magnetic flux acting from M1 to M2 and from M2 to M1. This means that the product of the cross-sectional area and the saturation magnetization Ms is approximately equal. For example, to form the domain wall conductive region M2, the saturation magnetization Ms is approximately equal to the cross-sectional area Ms. s (1) From the value of saturation magnetization M s (2) A material with a 40% higher value of θ can be used if the cross-sectional area of ​​this material is reduced by 40%. This is the case, for example, of Ni 81 Fe 19 This occurs when combining a domain wall path M1 of 1000 kA / m with M2 made of CoFe with a saturation magnetization of 1140 kA / m. This reduction can be achieved by reducing the width or thickness by 40%, or by combining both modifications, e.g. a 20% reduction in width and a 20% reduction in thickness. However, when reducing width and thickness, a symmetric reduction is preferred over a strongly asymmetric reduction. From the carried out micromagnetic simulations, it was found that as long as the average gap width is less than 50% of the layer thickness t of the magnetic conductor, B min and B max It can be assumed that the predicted value of will not change much. If the average gap width increases significantly, B min increases slightly, B maxIt is also expected to decrease slightly. When the average void width based on the layer thickness t changes by the maximum allowable amount, the width of the magnetic window decreases from 15 mT to 10 mT. By implementing the proposed invention, a spiral arrangement having a known structure suitable for counting the number of revolutions and located in one plane can be bridged using a structure that supports magnetic walls in a second plane and geometric interruptions, enabling the production of a novel closed loop. The present invention provides a series of possibilities that have not been achievable in the prior art for technical implementation, and if successfully implemented, it will lead to a significant expansion of the magnetic window compared to the solutions conventionally known in the prior art.

[0021] In the conventional transition section in the prior art (see, for example, DE102010022611B4), pinning of magnetic domains leading to the inoperability of the rotation counter is observed even for extremely small steps or lithography defects on the order of 15 nm, whereas the relatively large interruption of the magnetic wall conduction path at the location of the voids according to the present invention has no adverse effect on magnetic wall transfer.

[0022] All features apparent from the description, examples, claims and / or drawings may be essential to the present invention either alone or in any arbitrary combination with each other.

Explanation of Reference Signs

[0023] E1 Plane in which the loop-shaped magnetic wall conduction path is located E2 Connection plane K1, K2 Intersection A1, A2, A3, A4, A5, A6 Circles showing enlarged views w Width of the magnetic wall conduction path M1 D Diagonal of the intersection D exp (Realistic) enlarged value of D X Rotation axis of the permanent magnet 12 DW Magnetic wall 12 Permanent magnet 13 Arrow indicating the direction of the magnetic field B-B Cross-section M1 Magnetic wall conduction path of the loop-shaped structure M2 Second magnetic wall conduction path for bridging M s (1) , M s (2) Different saturation magnetizations t Thickness of the magnetic wall conduction path M1 201 Gap between M1 and M2 101 Separation layer 102 Synthetic antiferromagnet 103 Antiferromagnet S1 Non-magnetic layer vs Vertical offset between M1 and M2 hs Horizontal offset between M1 and M2

Claims

1. In a rotary counter using a magnetic wall conduction path that is arranged in a loop shape and is located substantially within a plane (E1) and closed, a connection region where the inner end and the outer end of the loop portion of the loop of the first magnetic wall conduction path (M1) are placed together is bridged by a second magnetic wall conduction path (M2) attached to both ends of the spiral magnetic wall conduction path via one gap (201) each. The gap (201) forms a local interruption in the magnetic wall conduction path (M1) to be closed. In this gap (201), when the magnetic wall (DW) moves from the first magnetic wall conduction path portion (M2 or M1), a stray magnetic field is generated within the first magnetic wall conduction path portion (M2 or M1), and a width is provided such that it causes nucleation of the magnetic wall (DW) within the magnetic wall conduction path portion (M2 or M1) following the gap (201) in the moving direction. The average width of the gap (201) is determined to be smaller than the thickness (t) of the magnetic wall conduction paths (M1, M2). The adjacent magnetic wall conduction path portions (M2 or M1) are covered by a non-magnetic layer (S1) in the gap region. A rotary counter characterized by this.

2. The magnetic wall conduction path portions (M1, M2) can have a horizontal and / or vertical offset (hs, vs) from each other as long as it is less than 25% of the thickness (t) or width (w) of the magnetic wall conduction path (M1). A rotary counter according to claim 1, characterized by this.

3. The product of the cross-sectional area and the saturation magnetization in each of the magnetic wall conduction paths (M1 and M2) is substantially equal. A rotary counter according to claim 1, characterized by this.

4. Soft magnetic materials having different saturation magnetizations are used for the first and second magnetic wall conduction path structures (M1 and M2), and the difference in saturation magnetization is preferably kept less than 40%. A rotary counter according to claim 3, characterized by this.

5. The non-magnetic layer (S1) is made of a diamagnetic, paramagnetic or antiferromagnetic material. A rotary counter according to claim 1, characterized by this.

Images