I lllll llllllll Ill lllll lllll lllll lllll lllll 111111111111111111111111111111111
`US007128988B2
`
`c12) United States Patent
`Lambeth
`
`(IO) Patent No.:
`(45) Date of Patent:
`
`US 7,128,988 B2
`Oct. 31, 2006
`
`(54) MAGNETIC MATERIAL STRUCTURES,
`DEVICES AND METHODS
`
`(75)
`
`Inventor: David N. Lambeth, Pittsburgh, PA
`(US)
`
`(73) Assignee: Lambeth Systems, Pittsburgh, PA (US)
`
`( *) Notice:
`
`Subject to any disclaimer, the term of this
`patent is extended or adjusted under 35
`U.S.C. 154(b) by 325 days.
`
`(56)
`
`References Cited
`
`U.S. PATENT DOCUMENTS
`
`6,097,579 A
`6,146,776 A
`6,248,416 Bl
`6,262,869 Bl
`
`8/2000 Gill ......................... 360/324.2
`11/2000 Fukuzawa et al. .......... 428/692
`612001 Lambeth et al. ........... 428/65 .3
`7/2001 Lin et al. ............... 360/324.11
`
`FOREIGN PATENT DOCUMENTS
`
`JP
`
`408204253 A
`
`8/1996
`
`(21) Appl. No.:
`
`10/415,757
`
`(22) PCT Filed:
`
`Aug. 29, 2002
`
`(86) PCT No.:
`
`PCT /US02/27327
`
`§ 371 (c)(l),
`(2), ( 4) Date: Aug. 29, 2003
`
`(87) PCT Pub. No.: W003/021579
`
`PCT Pub. Date: Mar. 13, 2003
`
`(65)
`
`Prior Publication Data
`
`US 2004/0058196 Al Mar. 25, 2004
`
`(51)
`
`Int. Cl.
`GllB 5166
`(2006.01)
`GllB 5170
`(2006.01)
`(52) U.S. Cl. ................................................... 428/831.2
`(58) Field of Classification Search ................ 428/828,
`428/831.2, 832, 815
`See application file for complete search history.
`
`OTHER PUBLICATIONS
`
`PCT/US98/23855 filed Nov. 9, 1998.
`
`Primary Examiner-Holly Rickman
`(74) Attorney, Agent, or Firm-Armstrong, Kratz, Quintas,
`Hanson & Brooks, LLP
`
`(57)
`
`ABSTRACT
`
`A thin film magnetic structure, magnetic devices, and
`method of producing the same, wherein (110) textured,
`symmetry broken body centered cubic or body centered
`cubic derivative crystalline structures epitaxially grown on
`hexagonal shaped templates, in the presence of a symmetry
`breaking mechanism is provided to promote oriented
`uniaxial magnetic properties from a series of successively
`deposited film layers, result in new oriented magnetic layer
`structures and microstructures and thus improved magnetic
`devices and device performance.
`
`41 Claims, 15 Drawing Sheets
`
`I a1,<100>
`+65.26°
`
`I
`I
`I
`
`---hex, <110>
`
`oo
`
`c2,<100>
`-65.26°
`
`TDK Corporation Exhibit 1001 Page 1
`
`

`
`U.S. Patent
`
`Oct. 31, 2006
`
`Sheet 1 of 15
`
`US 7,128,988 B2
`
`~
`Cl)
`s....
`::J
`0)
`
`·-LL
`
`I-a::
`<{
`a::
`0 -a:: a..
`
`0
`00
`......
`
`0
`- -1 0 \
`
`00
`0
`
`(0) l\6Jau3
`
`N
`0
`
`TDK Corporation Exhibit 1001 Page 2
`
`

`
`U.S. Patent
`
`Oct. 31, 2006
`
`Sheet 2of15
`
`US 7,128,988 B2
`
`Ms
`
`Mx
`...................... ---
`~2
`
`My
`
`Hx
`
`Hk
`
`Ms
`
`Hk
`
`Figure 2
`
`PRIOR ART
`
`TDK Corporation Exhibit 1001 Page 3
`
`

`
`U.S. Patent
`
`Oct. 31, 2006
`
`Sheet 3 of 15
`
`US 7,128,988 B2
`
`4
`"---
`
`Figure 3
`
`PRIOR ART
`
`TDK Corporation Exhibit 1001 Page 4
`
`

`
`U.S. Patent
`
`Oct. 31, 2006
`
`Sheet 4of15
`
`US 7,128,988 B2
`
`-q(cid:173)
`Q)
`!......
`:::J
`0)
`
`·-u...
`
`0
`-------~
`
`0
`
`00 -
`
`co
`~~•._..._,._~- ----~ .... _._, ........ ___ , ______ ------· __ ...... --· ----- -- . -
`
`. /
`
`°"
`
`0
`
`(e)o ~ ~3
`
`TDK Corporation Exhibit 1001 Page 5
`
`

`
`U.S. Patent
`
`Oct. 31, 2006
`
`Sheet 5of15
`
`US 7,128,988 B2
`
`I a1,<100>
`
`+65.26°
`
`. . . . . . · . . . . . . . • .
`
`------hex, <110>
`oo
`
`c2, <100>
`-65.26°
`
`I
`I
`I
`
`Figure 5
`
`TDK Corporation Exhibit 1001 Page 6
`
`

`
`U.S. Patent
`
`Oct. 31, 2006
`
`Sheet 6of15
`
`US 7,128,988 B2
`
`0
`-----t-~---+-~---l--~-----__..,~~-+--~-i~
`
`co
`Q)
`s,._
`
`:::J en ·-LL
`
`0
`
`co -
`
`0
`~~~--~~--~~~~.~
`
`~
`I
`Ir)
`N
`0
`
`!")
`
`ci
`
`I
`N
`0
`
`.,..
`.....
`ci
`
`-
`
`0
`
`TDK Corporation Exhibit 1001 Page 7
`
`

`
`U.S. Patent
`
`Oct. 31, 2006
`
`Sheet 7of15
`
`US 7,128,988 B2
`
`0
`.----,.-....~~--~~..-~--.~~-r-~~-.-~---,~
`
`0
`1 - - - - - - - - - , ~
`
`0
`•~~~--~~~~~--~~--~-~-- ~
`
`... ~. ···-···--...... --.......... _ ---·--·-..................... -..... _. __ ·-- -··-·- -· .. _.....
`
`.. ... -.... °'
`
`0
`
`.____.._._~~--'-~~""-~--'~~--~~--~--o
`0
`N
`0
`
`TDK Corporation Exhibit 1001 Page 8
`
`

`
`U.S. Patent
`
`Oct. 31, 2006
`
`Sheet 8of15
`
`US 7,128,988 B2
`
`0
`--~--~----.~~-.------~~""T"-~--~----.~
`
`----·---·-- ~
`
`0
`
`0
`
`00 -
`
`co
`Q)
`~
`:J
`
`C"'J ·-LL
`
`- · --··· ·-·-··-· - - · --·· ••... ·-··· ..•. ·- -··---·-
`
`. ·--··-·-· -
`
`0
`. ---- .. ---··· ·-·· C"I
`
`...__~_._~_...~~_..___.__._~~----~_..__~--o
`0
`N
`M
`0
`ci
`
`V"\ -0
`
`-
`
`0
`
`TDK Corporation Exhibit 1001 Page 9
`
`

`
`U.S. Patent
`
`Oct. 31, 2006
`
`Sheet 9of15
`
`US 7,128,988 B2
`
`Normalized Magnetization Vs. Hard Axis Field, Hx
`
`My-{\
`,I
`14
`·····---------------------------[ _______ ,._ _ __ \ ________________________________ _
`\
`!
`I
`!
`,
`I
`i
`1
`'
`
`II
`
`II
`
`I
`
`'
`I
`
`_______ 1___________ ------------'-------
`I
`'
`
`1
`
`0.5
`
`0
`
`--en
`_..... -x >-
`
`en
`~~
`""O >-
`en
`!I-
`ct! ct!
`IW
`
`22
`
`. I
`He= 0
`I :
`! I
`------·-------------------------------- ..... ___________________ ,. _____________________ _
`
`I
`
`-0.5
`
`13
`
`-1
`-500
`
`Figure 9
`
`0
`Hx
`
`500
`
`TDK Corporation Exhibit 1001 Page 10
`
`

`
`U.S. Patent
`
`Oct. 31, 2006
`
`Sheet 10 of 15
`
`US 7,128,988 B2
`
`0
`~~~~--~__....~--~--...-~-.--~--.~
`
`------· ___ ,,_,, _____________ ·--·-·-·····-·---··- --· ~
`
`0
`
`0
`00
`
`- CD
`
`0
`
`~
`Q)
`'-
`:J
`C)
`
`·-LL
`
`---·-··----·-- .. ··--·---·-·--------·----·--·----·-------·-···----------··-- -- °'
`
`0
`
`'--~--'-~-'-~--'-~---''--_._ ___ ~_,_~_.o
`0
`N
`0
`
`0
`
`-
`
`TDK Corporation Exhibit 1001 Page 11
`
`

`
`U.S. Patent
`
`Oct. 31, 2006
`
`Sheet 11 of 15
`
`US 7,128,988 B2
`
`0
`!----~-~~---...::--~--~~~~~-~~
`
`• ---· -
`
`0
`
`QO -
`
`0
`---~---·CJ\
`
`N
`d
`
`.....
`VI
`d
`
`-
`
`0
`
`TDK Corporation Exhibit 1001 Page 12
`
`

`
`U.S. Patent
`
`Oct. 31, 2006
`
`Sheet 12 of 15
`
`US 7,128,988 B2
`
`z
`
`r
`
`y
`
`x
`
`Figure 12
`
`TDK Corporation Exhibit 1001 Page 13
`
`

`
`U.S. Patent
`
`Oct. 31, 2006
`
`Sheet 13 of 15
`
`US 7,128,988 B2
`
`26') I ir.15
`
`I I
`I
`
`I
`
`+J
`CD
`O>
`s.... co
`I-
`I I I
`I I
`
`"O
`Q)
`..c
`CJ)
`
`16
`
`17
`Magnets
`
`18
`
`I
`
`Disk Substrate
`
`Figure 13
`
`TDK Corporation Exhibit 1001 Page 14
`
`

`
`U.S. Patent
`
`Oct. 31, 2006
`
`Sheet 14 of 15
`
`US 7,128,988 B2
`
`9=0
`
`Figure 14
`
`TDK Corporation Exhibit 1001 Page 15
`
`

`
`U.S. Patent
`
`Oct. 31, 2006
`
`Sheet 15 of 15
`
`US 7,128,988 B2
`
`Normalized Magnetizations Vs. Hard Axis Field, Hx (Oe)
`
`2 0 -----------
`
`0.5
`
`0
`
`-0.5
`
`Cl)
`
`-..
`~
`>.
`(/) ro w _..
`>.
`2
`~
`-..
`(/)
`~
`"'O
`I.... ro
`I _..
`x
`2
`
`.1a---~~~~====~~L__~~~_l__~~~--'
`-250
`500
`-500
`0
`250
`Hx (Oe)
`
`Figure 15
`
`TDK Corporation Exhibit 1001 Page 16
`
`

`
`US 7, 128,988 B2
`
`1
`MAGNETIC MATERIAL STRUCTURES,
`DEVICES AND METHODS
`
`BACKGROUND OF THE INVENTION
`
`2
`to achieve and maintain in a manufacturing process where
`many different desired material properties must be obtained
`simultaneously.
`
`Background for Oriented Soft Magnetic Films
`
`20
`
`25
`
`1. Field of the Invention
`This invention is directed to magnetic material structures,
`methods for making magnetic material structures and
`devices made from magnetic material structures.
`2. Description of the Prior Art
`As the home, office, transportation system, business place
`and factory become more automated and electronically
`connected, and as electronic devices and appliances such as
`computers, communication devices, wireless communica(cid:173)
`tion devices, electronic games, entertainment systems, per(cid:173)
`sonal data assistants, transportation vehicles, manufacturing
`tools, shop tools, and home appliances become more sophis(cid:173)
`ticated there is, and will be, an ever-increasing demand for
`higher performance and low cost electronic circuits, sensors,
`transducers, data storage systems and other magnetic
`devices which employ magnetic thin film materials. In order
`for these devices to remain competitive in the market place
`each product generation must be higher performing, unob-
`trusive and, usually, less expensive than the previous. Hence
`there are ever increasing demands for technical improve(cid:173)
`ments in the materials and structure of these devices.
`For all of these applications the magnetic material has an
`improved performance if the magnetic properties can be 30
`better controlled during the construction. Two familiar prop(cid:173)
`erties, which are sometimes considered to be intrinsic mag(cid:173)
`netic properties, are the saturation magnetization, Ms, and
`the magnetocrystalline anisotropy energy density constants
`(usually denoted by a subscripted K symbol). The meaning 35
`of magnetic anisotropy energy is that the magnetization
`would have a preferred direction, or directions, of orienta(cid:173)
`tion. That is, the energy of the system is minimal when the
`magnetization vector points along certain directions. These
`directions are referred to as the magnetic easy axes while the 40
`magnetic hard axes coincide with magnetic orientations
`where the energy is maximized. However, it should be noted
`that the magnetic anisotropy is not actually an intrinsic
`property in the sense that the materials are commonly not
`made perfectly. Nevertheless, good performance in device 45
`applications is almost always dependent upon there being a
`single preferred magnetic orientation or anisotropy direction
`and so in the manufacturing process one strives to achieve
`a desired uniaxial anisotropy. An objective of the present
`invention is to provide new mechanisms for controlling the 50
`magnetocrystalline anisotropy of thin magnetic films. By
`doing so the performance of almost all magnetic devices are
`envisioned to be improved.
`In general the anisotropy energy is a function of the
`orientation of the magnetization vector with respect to a 55
`given physical axis. Here, we define a "uniaxial" anisotropy
`to exist if the anisotropy energy density function only
`contains a single maximum and a single minimum as the
`magnetization angle, 8, is rotated by 180 degrees from a
`physical axis. Likewise we define an "ideal uniaxial" anisot- 60
`ropy energy to exist ifthe energy equation has only a sin2 (8)
`or cos 2 (8) dependence. Materials and device processing to
`achieve a desired orientation or anisotropy is commonly
`difficult and sometimes impossible, perhaps because here(cid:173)
`tofore the mechanism for achieving anisotropic orientation 65
`has not been well understood. Furthermore, uniform control
`of the orientation of the magnetic anisotropy is often difficult
`
`In magnetic devices, for example, such as sensors, trans(cid:173)
`ducers, transformers, inductors, signal mixers, flux concen-
`10 trators, recording media keepers, data recording and play(cid:173)
`back transducers it is common that the magnetic response to
`a driving field possess high sensitivity and at the same time,
`low coercivity (He). Or stated simply, the material possesses
`essentially non-hysteretic behavior. For this type of behavior
`15 the device is constructed so that the applied field is directed
`along the hard magnetic axis of a uniaxial magnetic material.
`This results in the minimization of coercivity and hysteric
`effects, which are many times associated with magnetic
`domain wall motion of materials that are multi-axial. For
`example, a material, which has bi-axial anisotropy, will have
`two easy and two hard magnetic axes and will exhibit
`hystersis and losses. In many of these applications a linear,
`or near linear response is also advantageous, while in other
`applications, such as signal mixers, a controlled non-linear
`response is desire. To obtain a linear magnetic response,
`requires both applying a field along the magnetic hard axis
`and that the anisotropy energy density function not only be
`uniaxial, but that it also have simple sin2 (8) or negative
`cos 2(8) dependence, where, 8, is the angle measured
`between the magnetization vector direction and the physi-
`cally determined magnetic easy axis. Since there is the
`mathematical identity, sin2(8)=1-cos 2(8), and since the ori(cid:173)
`gin in the energy function is arbitrarily defined the use of
`sin2(8) or -cos2 (8) yield equivalent physical behavior. Item
`[1] of FIG. 1 illustrates the squared sinusoidal anisotropy
`energy density curve shape versus the angle of the magne(cid:173)
`tization vector with respect to the easy axis located at zero
`degrees. FIG. 2 illustrates the response of the components of
`the magnetization, Mx and MY, as a function of applied field,
`Hx =Hm along the hard magnetic axis direction, x. The linear
`curve kinks only at the point [2] where the magnetization
`becomes saturated, or fully aligned with the applied field.
`For this special uniaxial anisotropy this occurs at the applied
`field value ofHk along the x direction, which is known as the
`anisotropy field. These response curves are sometimes
`referred to as hysteresis loops even though they exhibit no
`hystersis. It is the shape of the sin2(8) energy function that
`causes the response, Mx, along the hard axis to be linear and
`to be fully reversible. MY is the response in the y direction
`to an applied field in the x direction. The curve shape shown
`is quadratic for applied field magnitudes less than Hk, where
`MY is zero for larger magnitude fields. The quadratic behav-
`ior is necessary for linear Mx since M/=M/ +M/, where Ms
`is the total, constant, saturation magnetization vector mag-
`nitude. If the anisotropy energy is uniaxial, but is not
`governed by the, ideal, sin2 (8) functional form then the
`magnetic response is not linear. However, heretofore, the
`applicant knows of no real material examples exhibiting
`both a uniaxial energy curve and a non-linear Mx versus Hx
`behavior.
`Materials exhibiting the sin2 (8) energy density functional
`form are often referred to as having Stoner-Wohlfarth behav(cid:173)
`ior after the famous ideal uniaxial single domain magneti(cid:173)
`zation theory. However, thin films are commonly multi-
`the sin2 (8)
`domain even though
`they might exhibit
`
`TDK Corporation Exhibit 1001 Page 17
`
`

`
`US 7, 128,988 B2
`
`3
`functional energy form on a localized basis. Unless the hard
`axis direction is the same at all points in a sample and the
`applied driving field is exactly parallel to the hard axis then
`domain wall motion can commonly be observed. This
`motion results in coercivity mechanisms and hysteretic
`energy losses. The lossless behavior of samples represented
`by FIG. 2 is due to the magnetization rotating in response to
`the applied field rather than a response via domain wall
`motion. Multi-axis anisotropy materials always switch via
`wall motion and so suffer losses.
`It is also well known that, for soft uniaxial thin films, by
`first applying a field along the easy axis, and then by keeping
`a constant bias field in this direction, to eliminate 180 degree
`domain walls, one can force all of the material to appear to
`be single domain as the hard axis is then driven. Hence, 15
`because of this bias field, H6 =HY, in the easy axis direction
`the application of any finite Hx field along the hard axis can
`never quite drive the magnetization vector completely to the
`energy maximum [3] and the response will always be
`reversible and so lossless. This is not the case for materials 20
`with multiple anisotropy axes. For uniaxial materials the
`rotational response is key to many sensor devices and it is
`common in various forms of magnetoresistive sensors to
`provide a bias field along the easy axis by either applying a
`small field or by exchange coupling the magnetic sensor
`material to a hard magnetic material that has been so
`oriented to provide an effective bias field.
`For some sensor applications, such as anti-theft devices,
`and special electronic mixing circuit devices, soft, low loss,
`magnetic properties are desired simultaneously with a spe(cid:173)
`cific non-linear response. In these applications, the driving
`field has historically, and most commonly, been directed
`along an easy axis or in the direction of the lowest magnetic
`anisotropy energy. In this direction magnetic domain wall
`motion is usually significant. This domain wall motion 35
`commonly results in a highly non-linear response or even in
`strong hysteretic behavior.
`Certain anti-theft, article surveillance, article identifica(cid:173)
`tion or inventory control devices rely upon detecting har(cid:173)
`monic signals, which are generated by this non-linear behav(cid:173)
`ior or upon materials being driven in to saturation. One of
`many examples, of this type of surveillance system and tag
`is described in U.S. Pat. No. 3,747,086. This type of tag
`response has also been disclosed as enabling multiple bits of
`information to identify objects in U.S. Pat. No. 5,538,803
`Other article tag devices are based upon the magnetoelastic
`effect and mechanical resonance, where coupling exist
`between the magnetization and the mechanical strain in the
`material. An example, of this type of tag is disclosed in U.S.
`Pat. No. 4,510,489. In these later devices it is desirable to
`drive the magnetization towards a hard axis so that rotation
`of the magnetization dominates and magnetic hysteretic
`losses are minimized. By using this mode, and by using the
`field to drive the device at its mechanical resonance fre(cid:173)
`quency significant amounts of energy can be stored in the
`device. Hence, even after the drive field has been removed
`the mechanical vibration continues, the magnetoelastic
`properties are then used to invert the process to transmit a
`magnetic field as the mechanical stress causes the magne(cid:173)
`tization vector direction to rotationally oscillate synchro- 60
`nously with the mechanical vibration. This time dependent
`magnetic dipole radiates a magnetic field at the distinct
`resonance frequency, which can be detected to verify the
`tag's existence. Uniaxial anisotropy is needed in such a
`device to achieve low losses. Due to the magnetoelastic
`coupling between the mechanical strain and the magnetic
`moment orientation the fundamental of the mechanical
`
`4
`resonance frequency is emitted as an oscillating magnetic
`dipole field. However, ifthe uniaxial anisotropy is non-ideal
`and results in a non-linear response then the harmonics
`would also be available for detection in the presence of a
`drive signal. This is advantageous, but there have never been
`available materials from which to form such a device before.
`Likewise, in analog mixer circuit devices a non-linear
`response is desired. When two separate sinusoidal signals
`are simultaneously imposed on a nonlinear circuit device
`10 component a multiplication process results in a beating of
`the two signals. This results in additional harmonics at the
`sum and difference frequencies of the initial signals. Hence,
`the information contained in a modulated carrier frequency
`signal can be shifted to a beat frequency. Typically this is
`done to shift the information carrying bandwidth to a higher
`carrier frequency bandwidth (modulation) or bring the infor-
`mation carrying bandwidth signal back down to a bandwidth
`located near, or nearer to, zero frequency for demodulation.
`These techniques of frequency shifting are common to
`telecommunication and signal processing and in many other
`signal processing applications. The non-linear circuit
`response, circuit components, and circuit implementation
`used to perform these signal mixing processes are described
`in many electrical engineering circuit textbooks. Two
`25 example texts are "Electronic Communications Techniques"
`by P. H. Young and "Physics of Semiconductor Devices" by
`S. M. Sze. Ideally the non-linear device used to mix the
`signals is efficient, low loss, and low noise. Usually non(cid:173)
`linear Silicon active devices are used for this. In the past
`30 when magnetic devices were used for this application
`domain wall motion caused both losses to the signal and
`induced noise. A magnetic material with non-linear response
`that operates by low loss magnetization rotation rather than
`by domain wall motion is very desirable.
`For most all transducer and sensor applications, which
`require low anisotropy values in order to provide large
`sensitivity, the use of cubic crystalline materials are com(cid:173)
`mon. However, due to the three fold crystalline symmetry of
`cubic materials, achieving a single axis of anisotropy energy
`40 density, which is almost always essential to obtaining the
`desired low loss, low noise, magnetic properties, is difficult.
`Thin, or thick, film materials are commonly employed. For
`example many devices, such as data storage playback trans(cid:173)
`ducers or field transducers, commonly utilize face centered
`45 cubic, fee, thin film crystalline materials. Should these
`materials be prepared with a (001) crystalline texture the
`anisotropy energy, as a function of angle in the film plane is
`bi-axial, has multiple easy and hard axes in the film plane
`yielding a non-linear and hysteretic magnetic response
`50 resulting in noisy signals. Hence, a (111) crystalline texture
`is desired, where it can be mathematically shown for the
`case of the magnetization being confined to the (111) texture
`plane, that due to the crystalline three fold symmetry of the
`projection of the { 111} crystalline directions into the film
`55 plane, the cubic material will possess no net first order
`anisotropy energy density. Even for only moderate satura(cid:173)
`tion magnetization thin films, the magnetization is essen(cid:173)
`tially confined to the film plane by the demagnetization
`forces associated with the planar film shape.
`To achieve a single uniaxial anisotropy in a cubic material
`with (111) texture the material is usually subjected to a
`thermal treatment, or is deposited directly, in the presence of
`an applied magnetic field. The resulting single magnetic
`easy axis is aligned along the applied magnetic field, while
`65 the hard axis is perpendicular to this applied field. Further(cid:173)
`more the Mx versus Hx response function is a linear
`response. While the mechanism for the cause of this induced
`
`TDK Corporation Exhibit 1001 Page 18
`
`

`
`US 7, 128,988 B2
`
`5
`magnetic anisotropy is not well understood it is often argued
`that an atomic pair ordering mechanism occurs to break the
`symmetry of the directions in the (111) plane for each grain
`of the material. That is, on a local scale inside each grain,
`pairs of atoms align along the applied field during the
`deposition, or the annealing process, to reduce the magnetic
`field energy. Interestingly, annealing in a field of different
`orientation can often alter this induced anisotropy direction
`demonstrating that the orientation inducing mechanism is
`reversible. It is believed that this localized ordering allows 10
`fee materials such as the NiFe alloys, permalloy, to have a
`small induced uniaxial anisotropy. Magnetic thin film body
`centered cubic, bee, or bee derivative materials, such as Fe,
`FeCo, FeAI, and similar compositions, are seldom used in
`such applications because the required (111) crystalline 15
`texture never develops during deposition of the bee sym(cid:173)
`metric crystals.
`It is well understood that during the growth of thin metal
`films certain texture orientations tend to appear. These are
`driven by the minimization of surface energy and surface 20
`bonding considerations and are modified by surface mobility
`issues, which can be somewhat controlled by substrate and
`processing conditions. One simple rule of thumb is that the
`film surface energy is minimized when the atomic surface
`configuration is the most closely packed. For a fee crystal 25
`the atoms in the surface are most closely packed in the (111)
`plane and so this texture is the most likely. The (001) texture
`is less energetically likely, but is possible where as the high
`surface energy (110) texture never seems to occur. On the
`other hand, for a bee crystal the atomic surface most closely 30
`packed is the (110) texture and this commonly appears, the
`(001) texture is the next lowest energy and is sometimes
`induced, but the high surface energy (111) texture never
`seems to occur. Hence, for fee crystals we refer to the (111)
`texture as being natural where as for the bee crystal we refer 35
`to the (110) texture as being naturally occurring. If a low
`surface energy substrate is chosen, such as an amorphous
`metal, which tends to wet to the depositing material, the fee
`(111) texture and the bee (110) texture naturally occur. On
`the other hand, if the substrate is non-wetting, such as an 40
`oxidized surface where the depositing material tends to bond
`to the oxygen atoms to limit the atomic surface mobility,
`then it is common to see limited orientation in the deposited
`films or at best a set of mixed textures. The fee crystal tends
`to form only weak (111) and (OOl)textures while the bee 45
`crystal tends to form only weak (110) and (001) textures. For
`magnetic cubic crystalline thin film grains with (110) or
`(001) texture there are multiple easy or hard axes in the film
`plane of the grain. In addition, because each grain has a
`random in-plane orientation relative to other grains, these 50
`polycrystalline magnetic materials result in an ensemble of
`grains with multiple, random, anisotropy axes. These mate(cid:173)
`rials yield both a non-linear response, as well as, high
`coercivity associated with losses and noise. Hence, bee or
`bee derivative materials, which almost always grow with 55
`(110) or (001) texture, are seldom used for devices. For this
`reason the high saturation magnetization value of bee mate(cid:173)
`rials have largely been unavailable to the device designer
`who wishes to avoid domain wall motion.
`One need only consult the very authoritative text, "Phys- 60
`ics of Ferromagnetism, 2nd Edition" by Soshin Chikazumi,
`pages 299-309, on thermally induced anisotropy of permal(cid:173)
`loy to see that the degree of understanding of the cause of
`uniaxial anisotropy in cubic materials is poor. Professor
`Chikazumi details the literature on the subject and notes that 65
`one theory attempts to explain this "phenomena in terms of
`"directional order," or an anisotropic distribution of different
`
`6
`atomic pairs such as Ni-Ni, Fe-Fe, or Ni-Fe." The logic
`is that the atomic spacing between the Ni-Fe pair is smaller
`than the other possible pairs and so a lattice distortion results
`from the atomic pairing. A magneto-crystalline uniaxial
`anisotropy is then proposed to result from the magneto(cid:173)
`elastic energy associated with the resulting lattice distortion.
`Professor Chiakazumi also outlines a second theory in which
`it is "assumed that ordering occurred by the growth of
`distinct volumes of the ordered phase, and explains the
`induced anisotropy as the result of shape anisotropy of the
`second phase." Geometrical grain shapes, other than spheri(cid:173)
`cal, could generate considerable shape anisotropy energy. It
`should be pointed out however, that no physical evidence
`supporting either of these theories has been provided.
`Shape anisotropy is an intellectually comforting explana(cid:173)
`tion, as this phenomenon is quite observable in elongated
`magnetic particles such as used in magnetic particulate data
`storage tapes and bar shaped permanent magnets. In these,
`the magnetic easy axis is aligned with the longer dimension.
`However, while Professor Chiakazumi illustrates that a
`rather complex pair ordering model with dipole-dipole inter(cid:173)
`actions proposed by Neel can be used to qualitatively
`explain field induced anisotropy, there is no quantitative
`agreement and he comments "The reasons why quantitative
`agreement is not obtained may lie in the approximate
`estimates of various quantities and in failure to attain
`complete thermal equilibrium." Nevertheless, he comments
`regarding the texture of the materials that "magnetic anneal(cid:173)
`ing is most effective for, <111> annealing, less effective for
`<110>, and least effective for <100>." In this statement he
`is referring to the parallel orientation of the applied field and
`the crystalline directions during the annealing process.
`Very recently a publication; "Soft High Saturation Mag(cid:173)
`netization (Fe0 _7Co0 _3 ) 1_xNx Thin Films For Inductive Write
`Heads," IEEE Transactions On Magnetics, Vol. 36, No. 5,
`September 2000, by N. X. Sun and S. X. Wang, claimed to
`achieve orientation of the bee like material during deposition
`in an applied magnetic field. The publication appears to
`indicate that the Nitrogen content was necessary, induced
`strain into the films, and created a small grain structure as it
`shifted the angle of, and broadened, the (110) x-ray diffrac(cid:173)
`tion peak. They also indicated that a "significant amount" of
`a second crystalline magnetic phase, Fe4N, appeared in the
`films. As Chikazumi suggests perhaps this anisotropy behav(cid:173)
`ior is due to the strain or shape or pair ordering associated
`with the second phase. It is interesting to note that they
`obtained this orientation on an oxidized Si (100) textured
`substrate and by the applicant's standards the FeCoN (110)
`texture is weak. They went on to indicate that by sandwich(cid:173)
`ing the FeCoN film between two permalloy films the hard
`axis coercivity could be decreased, but they did not indicate
`or offer any evidence that the orientation of the FeCoN film
`was improved by depositing upon the permalloy. In fact, a
`comparison of hard axis hystersis loops for films prepared
`with and without the permalloy films indicate similar anisot(cid:173)
`ropy fields necessary to saturate the magnetization. This
`would indicate that the permalloy layer did not improve the
`orientation in their structure.
`
`Background for Orientation with Hard Magnetic
`Films
`
`It should be noted that unlike field sensing and energy
`transforming devices that usually require soft magnetic
`materials, devices such as magnetic recording media and
`
`TDK Corporation Exhibit 1001 Page 19
`
`

`
`US 7, 128,988 B2
`
`7
`permanent magnets demand higher uniaxial anisotropy
`energy in order to achieve high coercivity He, and a pre(cid:173)
`ferred orientation in order to achieve a high remnance value.
`For motors or actuators this directly affects the amount of
`work that a device can deliver while for recording media it
`directly affects the output signal level and signal pulse
`width, or flux transition width, and, hence, the recording
`density. Even in some of these hard magnet applications it
`is desirable to incorporate soft magnetic materials to
`enhance overall performance. For example, perpendicular 10
`thin film recording media has long been discussed as a future
`replacement for longitudinal thin film media. However, in
`order for a recording system to function properly it is
`desirable that a soft magnetic keeper layer, or underlay er, be
`placed on the opposite side of the perpendicular hard mag- 15
`netic recording layer from the recording head. This soft layer
`then provides a flux return path, or flux concentrator, for the
`recording head fields, as well as, a flux closure path for
`stabilizing the recorded bits against demagnetization ener(cid:173)
`gies after the head is removed. The former enables recording 20
`heads to function with higher coercivity media and with
`better resolution, while the later provides improved stability
`to the recorded patterns by eliminating a portion of the
`self-demagnetization energy associated with perpendicular
`recording. In longitudinal recording media a soft magnetic 25
`layer would also improve the stability of the recorded
`patterns by reducing the self-demagnetization energy of the
`hard recording layer. However, heretofore, no soft magnetic
`underlayer has been found to be a satisfactory keeper layer
`for either perpendicular or longitudinal media as no good 30
`method of controlled the anisotropy orientation has been
`available. Soft magnetic underlayers, without a uniform and
`controlled anisotropy direction, results in domain wall
`induced media noise via domain wall motion Barkhausen
`phenomena. As the magnetic head passes over the media 35
`data it effectively shorts out the stray media bit fringe fields
`causing the magnetic patterns in the soft magnetic under(cid:173)
`layers to relax to new positions. If there are significant
`domain walls involved in this process they commonly break
`loose from
`localized pinning defects causing sudden 40
`changes in their magnetic configuration. This Barkhausen
`phenomena causes noise signals to appear in the data
`playback head. For either perpendicular or longitudinal hard
`disk recording, where the recorded bit o