Keynote: Thin Film Magnetic Heads: Break-through of Electroplating into Electronics, Quantum Jump in Magnetic Storage

Tuesday, 7 October 2014: 14:00
Expo Center, 1st Floor, Universal 1 (Moon Palace Resort)
L. Romankiw (IBM TJ Watson Research Center) and S. Krongelb (Emeritus, IBM T.J. Watson Research Center)
When IBM announced the random access magnetic disk storage system RAMAC 305 in 1957, it was hailed as a major break-through in data storage. By the early 1960s engineers realized that within a decade or so the system would not be able to achieve the higher areal density needed to meet the ever increasing demand for greater storage capacity. The limitation was that the hand-wound magnetic head could not be made any smaller to further increase the areal density on the discs. Many proposals were made and patents filed for small heads, but the technology to produce the required miniature 3-D structures did not exist.

The first break-through came when we demonstrated that it is possible to electroplate the 3 µm thick, precise patterns of copper and permalloy required for a magnetic head through a photoresist mask on a thin, conductive seed layer. This was followed by exploratory investigations to achieve the required magnetic material properties of the metallic alloys and the invention of appropriate plating tools to build the head structures. Working side by side with the design engineers, functional requirements of the head could be balanced against fabrication process capabilities to produce a testable device.

Our presentation will describe the structure of the inductive read/write heads, MR read-only heads and integrated MR read/inductive write heads. We will show that viable heads would not have been possible without several studies and inventions relating to plating solution chemistry, processing, materials and very uniform, high intensity, reproducible agitation in plating tools.  Furthermore, all the process steps had to be carefully chosen to make sure no process step damaged any of the previously fabricated structure or interfered with subsequent steps. We were learning process integration before that concept even had a formal name.

Once the prototype heads were built at the Yorktown Research lab and accepted for IBM’s next generation storage system, the next step was to design and build a pilot line in San Jose, IBM’s manufacturing site for magnetic storage. Electrochemical processes with the precision required for the thin film head had never before been run on a production line. Establishing the line required frequent visits to San Jose and often daily phone contact with their engineers. Furthermore, several of the engineers who were hired to run the San Jose line spent their first year with IBM in Yorktown doing research related to head fabrication, thus facilitating the transfer of this knowledge to San Jose.

The five prong approach that prevailed from the first prototype to the final product was to concurrently develop the fabrication tools and to engage in studies related to the materials, the process, process control and process integration. Our preference was to use concrete data wherever possible, but to make the best engineering judgment if such data was not available. Studies were carried just far enough to obtain the information required to control the process and build the head. When time permitted, or if new problems surfaced, we could further pursue these studies. Where problems could be clearly defined and did not need head building capability, we tried to get professors and students interested in working on them.

We have demonstrated that even in the absence of a full set of concrete data it is possible to build a relatively complex device by using the available information in combination with the best engineering judgment and then coming back to refine the technology into a full-fledged manufacturing process with all the necessary controls for optimum mass production.

When the RAMAC 305 was first introduced, the disk size was 24 inches in diameter and the density of magnetic recording was 2 kilobits/in2. Today the disk size may be as small as two inches in diameter and the density is approaching a trillion bits per square inch – an increase of nearly 6 orders of magnitude in 50 years. In the same time, the cost of storage has dropped by 6 orders of magnitude. The successful use of electrochemistry in building the thin film head has underscored the importance of this field in both science and engineering. Electrochemical technology today is treated as a desirable fabrication approach in silicon chip fabrication, packaging, chip package attachment and MEMS. The thin film head has been a key factor in enabling the computers that now play a vital role in every aspect of our lives.