Not Applicable.
Not Applicable.
U.S. Pat. No. 5,640,343, Jun. 17, 1997—Gallagher et al.
U.S. Pat. No. 7,224,601, May 29, 2007—Panchula
U.S. Pat. No. 7,529,121, May 5, 2009—Kitagawa et al.
U.S. Pat. No. 7,440,339, Oct. 21, 2009—Nejad et al.
A magnetic random access memory (MRAM) using spin-induced switching is a strong candidate for providing a dense and fast non-volatile storage solution for future memory applications. An MRAM includes an array of memory cells arranged in rows and columns.
Cell size is one of the key parameters of MRAM. It substantially depends on the size and number of selection transistors supplying a spin-polarized write current to an MR element. The number of the transistors controlling the write current usually varies from one or two per MR element. It depends on the saturation current of the selection transistor and magnitude of the spin-polarized current required to cause switching of the MR element. Frequently, especially for MR elements having in-plane magnetization in the magnetic layers, one selection transistor cannot provide the required spin-polarized current due to its saturation. This obstacle prevents MRAM cell size reduction.
Another important parameter of the MRAM is a write speed. The write speed depends on the magnitude of the spin-polarized current running through the MR element. High speed (short duration of the write current pulse) requires higher magnitude of the spin-polarized current that can be limited by the saturation current of the selection transistor or by a breakdown of the tunnel barrier layer.
The present disclosure addresses to the above problems.
In accordance with one embodiment a magnetic memory device comprises: a substrate and a stack of planar memory arrays disposed on a substrate surface, each memory array comprising a plurality of parallel first conductive lines, a plurality of parallel second conductive lines overlapping the first conductive lines at a plurality of intersection regions, a plurality of magnetic tunnel junctions, each magnetic tunnel junction being disposed at an intersection region, electrically coupled to one of the first conductive lines at a first terminal and to one of the second conductive lines at a second terminal, and comprising a controllable electrical resistance, wherein the electrical resistance of said each magnetic tunnel junction is controlled by a bidirectional spin-polarized current running between the first and second terminals in a direction perpendicular to the substrate surface.
In accordance with another embodiment a magnetic memory device comprises: a substrate and a plurality of planar memory arrays stacked on the substrate, each memory array comprising a plurality of parallel first conductive lines, each first conductive line comprising a ferromagnetic cladding, a plurality of parallel second conductive lines overlapping the first conductive lines at a plurality of intersection regions, a plurality of magnetic tunnel junctions, each magnetic tunnel junction comprising a controllable electrical resistance, being disposed at an intersection region and electrically coupled to one of the first parallel conductive lines at a first end and to one of the second parallel conductive lines at a second end, wherein the electrical resistance of said each magnetic tunnel junction is controlled by a joint effect of a magnetic field and a bidirectional spin-polarized current applied simultaneously to said each magnetic tunnel junction.
In accordance with yet another embodiment a magnetic memory device comprises: a substrate, a plurality of planar memory arrays vertically stacked on the substrate, each planar memory array comprising a plurality of parallel first conductive lines, a plurality of parallel second conductive lines overlapping the first conductive lines in a plurality of intersection regions, a plurality of magnetic tunnel junctions, each magnetic tunnel junction being disposed at an intersection region and comprising a free ferromagnetic layer having a reversible magnetization direction, a pinned ferromagnetic layer having a fixed magnetization direction, and a tunnel barrier layer disposed between the free ferromagnetic layer and the pinned ferromagnetic layer, said each magnetic tunnel junction being electrically coupled to one of the first conductive lines at a first terminal and to one of the second conductive lines at a second terminal; a circuitry area disposed on the substrate beneath the plurality of planar memory arrays and comprising an electrical circuitry; and a plurality of interconnects disposed adjacent to the circuitry area to provide an electrical coupling of the plurality of first conductive lines and the plurality of second conductive lines to the electrical circuitry, wherein an electrical resistance of said each magnetic tunnel junction depends on a mutual orientation of the magnetizations directions in the free and pinned ferromagnetic layers.
12 pinned (or reference) magnetic layer
14 tunnel barrier layer
16 free (or storage) magnetic layer
20, 30, 60, 100, 120 magnetic random access memory (MRAM)
22 array of memory cells
24 bit line driver
26 word line driver
28 source line driver
52 conductor
54 magnetic flux concentrator
56 non-magnetic gap
1010 control circuitry area
1012 memory block
1014 bit line interconnects
1016 word line interconnects
1024 bit line interconnect contact area
1026 word line interconnect contact area
BL, BL1, BL2, BL3 bit line
C, C11-C33 memory cell
J, J11-J33 magnetic tunnel junction
SA1-SA3 sense amplifier
SL, SL1, SL2, SL3 source line
T, T11-T33 selection transistor
Tb1-Tb6 bit line transistor
Ts1-Ts3 read transistor
Tw1-Tw6 word line transistor
WL, WL1, WL2, WL3 word line
Embodiments of the present disclosure will be explained below with reference to the accompanying drawings. Note that in the following explanation the same reference numerals denote constituent elements having almost the same functions and arrangements, and a repetitive explanation will be made only when necessary.
Note also that each embodiment to be presented below merely discloses a device or method for embodying the technical idea of the present disclosure. Therefore, the technical idea of the present disclosure does not limit the materials, structures, arrangements, and the like of constituent parts to those described below. The technical idea of the present disclosure can be variously changed within the scope of the appended claims.
Refer now to the drawings,
The MR element herein mentioned in this specification and within the scope of the claims is a general term of a tunneling magnetoresistance (TMR) element using an insulator or semiconductor as the tunnel barrier layer. Although the above mentioned figures each illustrate the major components of the MR element, another layer (or layers) such as a seed layer, a pinning layer, a cap layer, and others may also be included.
Each memory cell comprises an MR element without a selection transistor. The MR element is connected to the appropriate bit and word lines at its ends and disposed at the intersection of the lines in a vertical space between them. A schematic view of the memory cell of the MRAM 30 is shown on
In the MRAM 30 shown in
The bit lines BL1-BL3 extend in the X-direction. They are electrically connected at one end to a bit line driver 24 that includes CMOS transistors Tb1-Tb6. For example, the bit line BL2 is connected at one end to a common drain terminal formed by an n-type transistor Tb3 and p-type transistor Tb4. A source terminal of the p-type transistor Tb4 is connected to a power supply. A source terminal of the n-type transistor Tb3 is connected to a ground. Similarly the bit lines BL1 and BL3 are connected to the pairs of CMOS transistors Tb1, Tb2 and Tb5, Tb6, respectively. Gate terminals of the transistors Tb1-Tb6 are connected to the bit line driver 24. The bit line driver 24 operates as a row selection switch.
The word line WL1-WL3 extend in the Y-direction crossing the X-direction. One end of each word line WL1-WL3 is connected to the word line driver 26. The driver 26 comprises a plurality of read/write circuits. Each of the read/write circuits includes at least a pair of CMOS transistors comprising one of p-type transistors Tw2, Tw4 or Tw6 and one of n-type transistors Tw1, Tw3 or Tw5 connected in series to each other, and one of a sense amplifiers SA1-SA3. Each of the transistors pairs Tw1 and Tw2, Tw3 and Tw4, and Tw5 and Tw6 is connected to a power supply at a source terminal of the appropriate p-type transistor and to the ground at a source terminal of the appropriate n-type transistor. The word line is connected to a common drain terminal of the CMOS transistor pair and to one input terminal of the sense amplifier SA through a read transistor Ts. For example, the word line WL2 is connected by its end to the common drain terminal formed by the transistor Tw3 and Tw4 and to the first input terminal of the sense amplifier SA2 through the read transistor Ts2. Second input terminal of the sense amplifier SA2 is connected to a reference element (not shown). Gates of the transistors Tw1-Tw6 are connected to the word line driver 26. The driver 26 operates as a column selection switch.
The sense amplifier SA1-SA3 comprises at least two inputs. One input of the amplifier is connected to the end of the word line WL1-WL3 and to the common drain terminal of the transistor pair by means of the read transistor Ts1-Ts3. The other input of the sense amplifier is connected to a reference element (not shown). The sense amplifier judges a data value of the MR element inside of the selected memory cell based on a reference signal Ref.
The memory 30 shown in
The MRAM 30 shown in
According to theory, the magnitude of the minimum spin-polarized current that is required to reverse the magnetization direction in the free layer is given by
where e is an electron charge, h is Planck's constant, α is Gilbert's damping constant, MS is saturation magnetization of the free layer material, V is volume of the free layer, and p is a spin polarization of the current. The factor g(θ) depends on the relative angle θ between vectors of magnetization (shown by arrows in
HEFF//=HK//+2πMS+HAPP+HDIP (2)
HEFF⊥=HK⊥−4πMS+HAPP+HDIP, (3)
where HK// and HK⊥, are the field of uniaxial crystalline anisotropy of in-plane and perpendicular magnetic material, respectively; HAPP and HDIP are the applied external field and the dipole field from the pinned layer acting on the free layer. The factor −4πMS arises from the demagnetizing field of the thin film geometry of the free layer having the perpendicular anisotropy. The same factor for the free layer with in-plane anisotropy is equal to +2πMS. Hence, the MTJ with perpendicular anisotropy may require substantially smaller (depends on HK and MS) switching current than that with similar parameters but having the in-plane anisotropy.
The direction of the magnetization in the free layer 16 of the MR element in its equilibrium states can be parallel or anti-parallel to the magnetization direction in the pinned layer. At these conditions the switching current that is required to reverse the magnetization direction in the free layer has its maximum value. Moreover, the magnitude of the current depends significantly on the duration of the current pulse. The magnitude of the switching current is almost inversely proportional to the pulse duration. Hence, the high speed writing (short current pulse) requires high switching current. The magnitude of the switching current is limited by the probability of a tunnel barrier layer breakdown. The above obstacles limit switching speed and endurance of MRAM with spin-induced switching.
The equation (1) suggests that the spin-polarized write current can be reduced significantly by changing the angle θ between the vectors of the magnetization in the free and pinned layers. Since the orientation of magnetization in the pinned layer 12 is fixed, the angle θ can be changed by tilting the magnetization in the free layer 16 from its equilibrium state. Tilting of the magnetization of the free layer 16 can be provided by applying a bias magnetic field along a hard magnetic axis of the free layer 16.
The bias magnetic field HB generated by the bias current IB is proportional to the current. For example, a current of 0.1 mA can generate a bias magnetic field of about 10 Oe in the vicinity of the MR element made with 65 nm technology node. This magnitude of the bias field HB is not sufficient to cause an unwanted reversal of the magnetization in the memory cells exposed to the bias field. The reversal of the magnetization can be achieved when both the bias magnetic field HB and spin-polarized current IS affect the selected MR element simultaneously. Hence the proposed hybrid writing mechanism provides a good selectivity of the MR elements in the array and significant reduction of the spin-polarized current IS. That is important for achieving a high endurance of MRAM operating at high speed, especially.
To write a logic “0” to the MR element J22 (
The input signals applied to the gate of the transistors Tb3, Tb4, and Tw4 are synchronized in time. Pulses of the currents IB and IS can overlap each other partially (shifted in time) or completely. The order of the pulses at partial overlapping can be any. The transistor Tb4 should be opened while any of the transistors Tb3 or Tw4 are opened.
The memory 60 also provides a possibility of simultaneous writing to the several MR elements having electrical contact with the energized bit line BL2 (
Data can be written to the memory cells C21, C22, and C23 at the same time by applying an appropriate signal to the gate of the transistors Tw1 or Tw2, Tw3 or Tw4, and Tw5 or Tw6. Simultaneous writing to several memory cells can provide significant reduction of a write energy per bit by means of more effective use of the bias current.
The transistors Tb1-Tb6 connected to the bit lines BL1-BL3 and the transistors Tw1-Tw6 connected to the word lines WL1-WL3 are exposed to different magnitudes of the current running through them during writing. Therefore they can have different saturation currents that can be achieved by using different size of transistors or by using several transistors. For instance the transistors Tb1-Tb6 can have higher saturation current than the transistors Tw1-Tw6. The transistors Tw1-Tw6 control the switching spin-polarized current in the MR elements of the array 22.
The MR elements of the disclosed MRAMs can use magnetic materials with in-plane and/or perpendicular direction of the magnetization in the equilibrium state. Magnetic materials exhibiting in-plane magnetization direction may have an uniaxial magneto-crystalline anisotropy that is not sufficient to provide a required thermal stability. To overcome this disadvantage the MR element employing in-plane magnetic materials (in-plane MR elements) can use a shape anisotropy in addition to the magneto-crystalline anisotropy. For that reason the MR element (or the free layer only) can have a shape of an elongated ellipse with an easy axis of the magneto-crystalline anisotropy oriented along a major axis of the ellipse. The ratio between the minor and major axis of the elliptical MR element can vary in the range from about 1:1.1 to about 1:5 depending on the technology node F of the manufacturing process. Frequently the technology node F can be equal to the size of the minor axis of the MR element. For example, the MR elements built using the technology node F=65 nm can have the ratio between the minor and major axis of about 1:2. It means, that lengths of the minor and major axis of the MR element are 65 nm and 130 nm, respectively.
Magnetic materials exhibiting a perpendicular magnetization direction usually have a substantial uniaxial magneto-crystalline anisotropy that can be significantly higher than that of the in-plane magnetic materials. For that reason the perpendicular MR elements having any shape, including round, can provide the required thermal stability.
MR elements made of magnetic materials with a perpendicular magnetization direction (anisotropy) are shown in
The arrays 22 of the memory cells shown in
The memory block 1012 includes four layers of the memory cell arrays 22-1, 22-2, 22-3, and 22-4 stacked one above another in the vertical direction (Z-direction) above a semiconductor substrate (or wafer). The layers of the memory cell arrays are electrically isolated from each other by insulator layers (not shown). Note that the number of the memory cell arrays in the block 1012 can be any. Each of the memory cell arrays 22-n comprises magnetic memory cells C arranged in rows and columns. The memory cell C can use a MR element J employing magnetic materials with in-plane (
As shown in
Connections of the bit lines BL and the word lines WL in the stacked memory cell arrays 22-n to the drivers 24 and 26, respectively require vertical conductive lines (via-contacts or interconnects) located on the sides of the memory block 1012. The bit lines BL can have one end connected to the bit line driver 24 by means of the vertical bit line interconnects 1014 that can be formed in a bit line interconnect contact area 1024. Respectively, the word lines WL can have one end connected to the word line driver 26 by vertical word line interconnects 1016 formed in a word line contact area 1026. The interconnect contact areas 1024 and 1026 can be provided along two sides of the control circuit area 1010. For example, the bit line interconnect contact area 1024 can be located along a side that is parallel to Y-direction, and the word line interconnect contact area 1026 can be located along a side of the control circuit area 1010 that is parallel to X-direction.
The configuration of the bit lines BL and word lines WL shared by the cell arrays 22-1-22-4 in the memory block 1012 and the relation between the bit line interconnects 1014 and the word line interconnects 1016 are described with reference to
As shown in
Selection of a memory cell (MR element) in the memory block 1012 can be achieved by applying appropriate signals to one bit interconnect line 1014 and to one word interconnect line 1016. The selected memory cell is located at the intersection of the activated bit and word lines. An operation of the 3D-memory 100 is similar to the operation of the 2D-memory 30 disclosed above (
During writing a logic “0” to the selected MR element J22 (
During writing logic “1” (
In the given exemplary embodiment the bit lines BL can be independently driven on a layer basis while the word lines WL can be commonly connected in all the layers (one line per layer). Note that the word lines WL may also be independently driven on a layer basis. Alternatively, the bit lines BL may be commonly connected while the word lines WL are independently driven. The bit lines BL and the word lines WL may be configured such that at least one of them is shared by the upper and lower layers. In this case, the upper and lower memory cells in the arrays 22-n are arranged symmetric about the common line. The arrangement of the MR elements J is not limited to the shown example.
As in the 3D-memory 100, the bit lines BL in the memory 120 can be independently connected to the bit line driver 24 by means of the appropriate vertical bit line interconnects 1014 (
During writing a logic “0” to the selected MR element J22 (
While the specification of this disclosure contains many specifics, these should not be construed as limitations on the scope of the disclosure or of what may be claimed, but rather as descriptions of features specific to particular embodiments. Certain features that are described in this specification in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable sub-combination. Moreover, although features may be described above as acting in certain combinations and even initially claimed as such, one or more features from a claimed combination can in some cases be excised from the combination, and the claimed combination may be directed to a sub-combination or variation of a sub-combination.
It is understood that the above embodiments are intended to be illustrative, and not restrictive. Many other embodiments will be apparent to those of skill in the art upon reviewing the above description. The scope of the embodiments should be, therefore, determined with reference to the appended claims, along with the full scope of equivalents to which such claims are entitled.
While the disclosure has been described in terms of several exemplary embodiments, those skilled in the art will recognize that the disclosure can be practiced with modification within the spirit and scope of the appended claims. Specifically, one of ordinary skill in the art will understand that the drawings herein are meant to be illustrative, and the spirit and scope of the disclosure are not limited to the embodiments and aspects disclosed herein but may be modified.
This application is a continuation-in-part of U.S. patent application Ser. No. 13/441,841, filed on Apr. 7, 2012 and claims the benefit of U.S. provisional patent application No. 61/472,788, filed on Apr. 7, 2011 by the present inventors.
Number | Name | Date | Kind |
---|---|---|---|
5640343 | Gallagher et al. | Jun 1997 | A |
6211090 | Durlam et al. | Apr 2001 | B1 |
6256223 | Sun | Jul 2001 | B1 |
6269018 | Monsma et al. | Jul 2001 | B1 |
6363000 | Perner et al. | Mar 2002 | B2 |
6801451 | Tran et al. | Oct 2004 | B2 |
6882553 | Nejad et al. | Apr 2005 | B2 |
6950333 | Hiramoto et al. | Sep 2005 | B2 |
7006375 | Covington | Feb 2006 | B2 |
7224601 | Panchula | May 2007 | B2 |
7313013 | Sun et al. | Dec 2007 | B2 |
7411817 | Nozieres et al. | Aug 2008 | B2 |
7432574 | Nakamura et al. | Oct 2008 | B2 |
7440339 | Nejad et al. | Oct 2008 | B2 |
7471550 | Koga | Dec 2008 | B2 |
7480173 | Guo et al. | Jan 2009 | B2 |
7502249 | Ding | Mar 2009 | B1 |
7508042 | Guo | Mar 2009 | B2 |
7511991 | Saito et al. | Mar 2009 | B2 |
7532505 | Ding | May 2009 | B1 |
7545672 | Ueda et al. | Jun 2009 | B2 |
7577021 | Guo et al. | Aug 2009 | B2 |
7596015 | Kitagawa et al. | Sep 2009 | B2 |
7652913 | Sugiyama et al. | Jan 2010 | B2 |
7668005 | Ueda | Feb 2010 | B2 |
7742328 | Chen et al. | Jun 2010 | B2 |
7755932 | Ito et al. | Jul 2010 | B2 |
7869265 | Shimizu et al. | Jan 2011 | B2 |
7924607 | Yoshikawa et al. | Apr 2011 | B2 |
8036010 | Maejima | Oct 2011 | B2 |
8149606 | Maejima | Apr 2012 | B2 |
8164940 | Ozatay et al. | Apr 2012 | B2 |
20090103354 | Yoon et al. | Apr 2009 | A1 |
20120107964 | Ranjan et al. | May 2012 | A1 |
Entry |
---|
X. Zhu and J.-G. Zhu, IEEE Trans.Magn., v.42, p. 2739-2741 (2006). |
S. Mangin et al., Applied Physics Letters, v.94, 012502 (2009). |
US 7,529,121, 5/2009, Kitagawa et al. (withdrawn). |
X. Zhu and J.-G. Zhu, IEEE Trans.Magn., v.42 p. 2730-2741 (2006). |
S. Mangin et al., Applied Physics Letters, v.94 012502 (2009). |
Number | Date | Country | |
---|---|---|---|
20120281465 A1 | Nov 2012 | US |
Number | Date | Country | |
---|---|---|---|
61472788 | Apr 2011 | US |
Number | Date | Country | |
---|---|---|---|
Parent | 13441841 | Apr 2012 | US |
Child | 13554467 | US |