Introduction to Micromagnetic Recording Physics
Extremely effective reduction of physical dimensions in the recording head through photolithography techniques, in combination with tremendous development in thin film recording media, has led to a spectacular increase in areal density in the magnetic rec
- PDF / 4,089,651 Bytes
- 35 Pages / 439 x 666 pts Page_size
- 21 Downloads / 215 Views
Recording Head Design
Michael Mallary
11.1
Introduction
Magnetic recording is literally as old as the hills. Recently a satellite, flying high over the polar regions of Mars, detected stripes of magnetisation spaced at more than 50 miles. This is proof that Mars had a significant magnetic field and a liquid core in the distant past. In order to detect finer structures, the height of the orbit would have to be reduced. At fly heights, h, greater than one third of the plus to minus spacing of the stripes, Sb, the field intensity from a periodic source falls as:
B ex: e -7rh/ Sb
(11.1)
This is known as the Wallace spacing loss formula. At this time the fly height in disk drives is about 25 nm. At this height the spacing loss factor is about 1/5 for a 50 nm bit space. This is approaching the minimum that a modern partial response maximum likelihood (PRML) channel detector can cope with, for a reasonable minimum signal to noise (SNR) requirement. The fly height in a disk drive is the key variable from which all the others scale. Recently density has doubled each year. This has been achieved by scaling all of the salient system dimensions, including tolerances, down by 30% per year.
11.2
Superparamagnetic Limit
At this time the recorded bits in all commercial media consist of a collection of several hundred randomly located switching units. Maintenance of adequate signal to noise requires that the size of these switching units (and therefore their total number per bit) be scaled with the size of the bits. Therefore the volume of the switching units, V, has been going down at 65% per annum. For decoupled grains, the energy barrier to thermally induced demagnetisation, E, is proportional to the volume (e.g., E = Ku V, Ku is the magnetic anisotropy energy per unit volume). For constant K u , E is also going down at 65% per year. In the near future, we will not be able to continue scaling down the size of these switching units. If their switching energy barrier falls below 40 kT they will not retain the written magnetic state for the requisite ten years (see M.L. Plumer et al. (eds.), The Physics of Ultra-High-Density Magnetic Recording © Springer-Verlag Berlin Heidelberg 2001
11
Recording Head Design
315
Chap. 5 for more detail on this). Approximately every nanosecond, thermal agitation attempts to switch grains' magnetic orientation. The probability that such an event will occur per attempt is: Pswitch
= e- E / kT
.
(11.2)
Here T is the temperature in degrees Kelvin, and k is Boltzman's constant. In ten years there will be e40 attempts (1 billion per second), hence the need for E > 40 kT. This is known as the superparamagnetic limit for magnetic recording. The superparamagnetic effect sets a lower limit to the switching volume below which we must not go. With scaling as usual, this would bring the growth of areal density to a screeching halt. However there are several ways to work around this problem which will be exploited in the near future. They require adjustments in the projected evolution of head a
Data Loading...
