Qualcomm’s Snapdragon S4 chipset is certainly hot (well, not too hot), but it looks like demand is expected to grow even further, causing the San Diego-based SoC maker to turn to allies in the east to help beef up supply. According to China Economic News Service, United Microelectronics Corp. (UMC) and Korea-based Samsung will join Taiwan Semiconductor Manufacturing Co. (TSMC) to manufacturer the 28nm chips beginning later this year, in an attempt to increase S4 availability ahead of the Windows RT launch. The article cites Qualcomm CEO Paul Jacobs as saying that a shortage is expected to continue, due to the complicated techniques necessary to manufacturer 28nm chips, and that the company may consider adding its own manufacturing plant in the future. All in all, it doesn’t seem like a terrible position for QCOM to be in. Full details are at the Taiwanese source link below.

There’s no question that Micron has shifted its focus away from PCs in favor of producing components, shipping everything from SSDs to CMOS sensors in recent years, but the semiconductor manufacturer just took a $2.5 billion step even closer to bridging its gap between other companies in the same market, including Samsung, the chip producer’s top competitor. Under the deal, Elpida Memory, which is headquartered in Tokyo, will fall within the Idaho-based conglomerate’s growing umbrella, netting Micron a 50-percent boost in production capability. That increase did come at great expense, however – the transaction included $750 million in cash and $1.75 billion in future installments (1,750 easy payments of one million dollars?), which are set to continue through 2019. The acquisition was also paired with a 24-percent stake in Rexchip Electronics for an additional $334 million, which will complement Elpida’s investment, yielding a total 89-percent stake for Micron. While the amount does seem quite significant, investors appear to be on board, with Micron’s stock ($MU) currently up more than 4 percent since this morning. Both deals will reportedly close within the next year.

A fundamental challenge of developing spintronics, or computing where the rotation of electrons carries instructions and other data rather than the charge, has been getting the electrons to spin for long enough to shuttle data to its destination in the first place. IBM and ETH Zurich claim to be the first achieving that feat by getting the electrons to dance to the same tune. Basing a semiconductor material on gallium arsenide and bringing the temperature to an extremely low -387F, the research duo have created a persistent spin helix that keeps the spin going for the 1.1 nanoseconds it would take a normal 1GHz processor to run through its full cycle, or 30 times longer than before. As impressive as it can be to stretch atomic physics that far, just remember that the theory is some distance from practice: unless you’re really keen on running a computer at temperatures just a few hops away from absolute zero, there’s work to be done on producing transistors (let alone processors) that safely run in the climate of the family den. Assuming that’s within the realm of possibility, though, we could eventually see computers that wring much more performance per watt out of one of the most basic elements of nature.

Intense semiconductor competition has already forced numerous Japanese companies to work together, and now Fujitsu has announced that it’ll merge its LSI chip design and R&D divisions with Panasonic. The two companies are looking to the state-run Development Bank of Japan to fund the new venture, which comes in the wake of expected Fujitsu losses of over $1 billion this year – forcing the company to cut 5,000 jobs and transfer 4,500 to other divisions by March 31st. Fujitsu said it’s also looking to transfer a state-of-the-art LSI fabrication line in central Japan to a new foundry venture with Taiwan Semiconductor Manufacturing, the world’s largest chip maker. That carries on a trend in declining Japanese chip dominance, exemplified by Elpida’s bankruptcy and the recent government bailout of Renesas, which itself is a merger of NEC, Hitachi and Mitsubishi’s semiconductor operations.

[Image credits: Wikimedia commons]

Harnessing the power of the sun is a tricky business, but even the past few weeks have seen some interesting developments in the field. In this latest installment, researchers from the Lawrence Berkeley National Laboratory and the University of California have figured out a way of making solar cells from any semiconductor, potentially reducing the cost of their production. You see, efficient solar cells require semiconductors to be chemically modified for the current they produce to flow in one direction. The process uses expensive materials and only works with a few types of semiconductors, but the team’s looking at using ones which aren’t normally suitable – the magic is to apply an electrical field to them. This field requires energy, but what’s consumed is said to be a tiny fraction of what the cell’s capable of producing when active, and it means chemical modification isn’t needed.

The concept of using a field to standardize the flow of juice isn’t a new one, but the team’s work on the geometrical structure of the cells has made it a reality, with a couple of working prototypes to satisfy the skeptics. More of these are on the way, as their focus has shifted to which semiconductors can offer the best efficiency at the lowest cost. And when the researchers have answered that question, there’s nothing left to do but get cracking on commercial production. For the full scientific explanation, hit up the links below.