April 2005

The Magnetic Race-Track Memory, a new concept in magnetic non-volatile memory disclosed by IBM Corp of the US in February 2005, is attracting a great deal of attention. If it can be commercialized as advertised it has the potential to revolutionize the memory architecture for computers and consumer electronics. The high potential performance of the new memory is the key, delivering performance on a par with dynamic random-access memory (DRAM) at the manufacturing cost of hard disk drives (HDD). It supports random access, and offers an infinite number of rewrites. Because no mechanism is required, it can be expected to deliver the same robustness as semiconductor memory.

But practical application is still "five to ten years away," said Stuart SP Parkin, IBM fellow SpinApps and director of IBM Research Division, Almaden Research Center at IBM.

The mean access time is about 50ns, which is a little longer than MRAM and roughly the same as existing DRAM.

Total revenue for the fourth quarter of fiscal 2005 was $3.07 million compared to $3.21 million for the prior-year quarter, a decrease of 4%. Product sales for the quarter increased 3% over the prior-year quarter and 42% compared to the third quarter of fiscal 2005.

Net income for the fourth quarter was $476,114 or $0.10 per diluted share, compared to $789,670 or $0.16 per diluted share for the prior-year quarter. Net income for the fourth quarter of fiscal 2005 included a net income tax benefit of $126,904 or $0.03 per diluted share, compared to $236,500 or $0.05 per diluted share for the fiscal 2004 quarter, from reductions of valuation allowances relating to deferred tax assets.

MRAM is an exciting opportunity; however even once you have working spin transistors there are other problems that must to be overcome before efficient MRAM can be produced. Ironically, one of the biggest problems is actually reading data from individual components of MRAM memory. The problem might be compared to trying to read a page of small text with a large magnifying glass, where you can only read one character at a time and the image is blurry with the characters on either side making it difficult to see the central character clearly.
The time taken to read a page of text this way makes the whole process unworkable, and it is the same with the “reader circuitry” currently used with MRAM. One alternative would be to shrink the magnifying glass down to the size of an individual character, but this would make identifying each character difficult, and in the same way when we try to shrink the reader circuitry for MRAM we find it no longer works!
The second breakthrough discovery made in Oxford solves this problem! It is a new type of reader circuitry that is simple, accurate and works quickly. The value of each component of MRAM can be easily read without any interference from adjacent cells, and the reader circuitry itself can be miniaturised down to the same scale as the individual units of memory. It works like having a line of lenses across a page, each the same size as an individual character, and allowing an entire line of text to be read instantly before moving on to the next one!

Read more here 

Magnetic Solutions today announced that it has accepted an order for an MRT2000 high field annealing oven for MRAM development and production. The order was placed by an international semiconductor company, that is a new entrant into the MRAM space.

The order was the result of a competitive evaluation between existing magnetic annealing tool manufacturers. The Magnetic Solutions MRT2000 was selected as the best technical solution for batch annealing of MRAM parts in production conditions and has a clear roadmap to 300mm production readiness.

Professor Hideo Ohno at the Tohoku University Research Institute of Electrical Communication and a research group of Hitachi's Advanced Research Laboratory have jointly developed a tunnel magneto-resistance (TMR) element with a reluctivity of 287%, the highest room temperature value ever measured.

The high reluctivity TMR element is necessary to speed up and reduce power consumption of MRAM, a nonvolatile memory that is regarded as a future super-gigabit memory candidate.
The new device also achieved 403% at low temperature.