Manchester MicroMagnetic Media Model

1
Future Magnetic Storage Media Jim
Miles Electronic and Information Storage Systems
Research Group
2
Future Magnetic Storage Media

1. Media requirements for very high density
2. Model description
3. Predicted effects of grain size distribution
4. Patterned media possible routes
5. Conclusions

3
Granular or Patterned Media?
4
(No Transcript)
5
Granular Media Limitations
6
Writing to Media
A sufficiently large field is needed to overcome
the anisotropy of the material, which keeps
magnetisation aligned along one axis
Anisotropy Ku
Magnetisation MS
7
Thermal Stability of Media

• Energy barrier EB KUV
• Thermal energy KBT
• Spontaneous switching
• when EB lt 70KBT
• Require EB 70 KBT

8
To Increase the Density

• Decrease the bit length Jitter must decrease
• Decrease the track width W Jitter must not
  increase.
• Jitter , ? grain diameter D must
  fall
• Volume V ?D2t/4 ? Volume falls
• ? KU must rise to keep EB KUV high enough
• ? bigger write field H gt 2KU/?0MS is needed.
• Density can only rise by increasing write field.

9
Perpendicular Recording
Increases write field, but only by x2
10
Other Problems of Granular Media

• Media are granular.
• Grains are not equal-sized.
• Typically ?D 0.2ltDgt, ?V 0.4ltVgt
• Hypothesis - Irregularity in media structure
  produces noise
• Big grains give big transition deviations
• Different grain volumes switch more or less
  easily
• Different grains see different local interaction
  fields.

11
Perpendicular Media Modelling
Real Storage Medium Model Storage
Medium (not to identical
scale)
12
Manchester MicroMagnetic Multilayer Media Model
(M6)
Mmmmmm...

• Landau-Lifshitz dynamic and M-C thermal solvers.
• Arbitrary sequences of uniform vector applied
  fields
• Recording simulation with FEM or analytical head
  fields.
• Soft underlayer by perfect imaging
• Microstructural clustering and texturing.
• Fully arbitrary grain positions and shapes.
• Full account of grain shape in interaction fields
• Allows vertical sub-division and tilted columns
  (MET like)

13
Magnetostatic Interaction - Pairs of Grains
Interaction Field Hj Dij Mi
Magnetostatic interaction tensors D are computed
numerically
Field grain experiences a field that varies
through the volume.
Surface charge from each polygon face of the
source generates field. Typically 48 faces per
polygon. Top and bottom faces computed similarly
by division into strips.
Integrate over the surface charge of i and the
volume of j. Underlayer included by incorporating
images into Dij
14
Exchange Interaction - Pairs of Grains
Exchange interaction factors are computed
numerically
Grain j experiences an exchange field due to
grain i
Integral term computed numerically from polygon
geometry
15
Varying Grain Size

• Voronoi seed positions randomised
• Minimum grain boundary width 0.7nm fixed
• Number of grains/m2 and packing fraction fixed
• Mean grain volume remains constant
• ? Hex remains constant

sv/ltvgt 0 sv/ltvgt 15
sv/ltvgt 39
16
Grain Size Distributions
sv/ltvgt 0 sv/ltvgt 4.7 sv/ltvgt 10.2 sv/ltvgt
15.5 sv/ltvgt 22.6 sv/ltvgt 29.4 sv/ltvgt
38.7
17
Exchange Field Distributions
Average exchange field does not change as the
microstructure changes. HE 0.5 HD A
1.85x10-13 for all structures
sv/ltvgt 0 sv/ltvgt 4.7 sv/ltvgt 10.2 sv/ltvgt
15.5 sv/ltvgt 22.6 sv/ltvgt 29.4 sv/ltvgt
38.7
18
Exchange Interaction Between Pairs of Grains
Width of line ? Hex
Uniform grains, perfect hexagonal lattice.
Exchange field is identical between all
pairs. Thermally decayed from DC saturated sv/ltvgt
0 HE/HD 0.5
19
Exchange Interaction Between Pairs of
Grains Width of line ? Hex
Large volume distribution sv/ltvgt 39 Irregular
structure, Large variation in HE ltHEgt/ltHDgt 0.5
20
Magnetostatic (Demag) Field Distributions
sv/ltvgt 0 sv/ltvgt 4.7 sv/ltvgt 10.2 sv/ltvgt
15.5 sv/ltvgt 22.6 sv/ltvgt 29.4 sv/ltvgt
38.7
21
Energy Barrier Distributions
sv/ltvgt 0 sv/ltvgt 4.7 sv/ltvgt 10.2 sv/ltvgt
15.5 sv/ltvgt 22.6 sv/ltvgt 29.4 sv/ltvgt
38.7
22
Recorded Transitions, b20nm, Tp 80nm, 411
Gb/in2
sv/ltvgt 39
sv/ltvgt 0
23
Effect of Irregularity on Data Signal
sv/ltvgt 0 sv/ltvgt 4.7 sv/ltvgt 10.2 sv/ltvgt
15.5 sv/ltvgt 22.6 sv/ltvgt 29.4 sv/ltvgt 38.7
24
Effect of Irregularity on Noise
sv/ltvgt 0 sv/ltvgt 4.7 sv/ltvgt 10.2 sv/ltvgt
15.5 sv/ltvgt 22.6 sv/ltvgt 29.4 sv/ltvgt
38.7
25
Grain Microstructure Conclusions

• Grain size distributions give rise to decreased
  signal and increased noise (BAD)
• Media with small grain size distributions are
  needed
• Patterned media are needed
• Additional advantage switching volume is the bit
  size, not the grain size ? lower switching field
  is possible.

26
(No Transcript)
27
(No Transcript)
28
Direct Write e-beam

1. Form master by direct write e-beam on resist
   layer
2. Evaporate gold coating
3. Lift-off gold from unexposed areas
4. Etch to remove magnetic layer except where
   protected by gold

50 nm diameter islands B. Belle et.
al. University of Manchester
29
Patterned Media Potential

• Provides a route to regular arrays of thermally
  stable low noise
• 1Tb/in2 requires 12.5nm lithography
• Not feasible using semiconductor manufacturing
  technology for some years to come

30
Self-Organised Magnetic Assembly (SOMA Media)

1. FePt nanoparticles manufactured in aqueous
   suspension.
2. Very narrow size distribution.
3. Deposited onto substrate.
4. Self-Assemble into ordered structure.

31
FePt Particle Growth
32
FePt problems
FePt manufactured in solution has low Ku. Very
high Ku can be developed by annealing
Much ongoing research in low temperature
formation of high coercivity FePt
33
Other Potential Technologies
Electro-chemical deposition in self-ordered
templates University of Southampton.
Electroplating into self-ordered pores in
Alumite R. Pollard et. al, Queens University
Belfast.
Vacuum deposition through self-assembled
nanosphere templates Paul Nutter, Ernie Hill,
University of Manchester.
34
Self-Assembly Long Range Order
Self-assembly produces only local order. Over
long ranges order breaks down at dislocations.
Self-assembled pattern using a diblock co-polymer
(in nanoimprinted grooves. (C. Ross et al, MIT,
2002)
40nm diameter CoCrPt nanoparticles. Mask made
from a diblock co-polymer (polystyrene/PMMA),
self-assembled in nanoimprinted grooves. (Naito
et al, Toshiba, IEEE Trans. Magn 38 (5) (2002)
35
Conclusions

• Conventional media can only be extended so far.
• Patterned media overcome thermal stability
  issues.
• Higher stability granular materials could be used
  with heat assisted recording (HAMR)
• but patterned media might still be needed to
  avoid excessive transition noise.
• Patterned media are likely to be necessary in 5
  years