News & Analysis
The ups and downs of three-dimensional circuits
Peter Clarke
12/20/2004 3:08 PM EST
LONDON — The possibility of integrating active circuits, and not just interconnect, above the wafer surface has been a goal for many years, but for a variety of reasons has not proved desirable.
One reason is cost and generally there is more than enough scope to design useful chips as planar circuits and let someone else, such as the system integrator, worry about the "z" dimension.
However, 2004 saw progress by two companies using two different methods to explore the third dimension and one company more or less forced to give up on a 3D research deal with Intel Corp.
In May Opticom ASA, a Norwegian research company with a Swedish subsidiary called Thin Film Electronics AB, prepared to break with Intel over its multi-layer plastic memory. Opticom had been working with Intel but announced in its interim results that technical issues related to mass production had not been solved and that TFE would not continue working with Intel.
Meanwhile, heavily-backed startup Matrix Semiconductor Inc. (Santa Clara, Calif.) lived up to a promise made a year before, and revealed it had shipped a 512-Mbit antifuse-based one-time-programmable memory to its lead customer in 2003. The technology uses up to eight layers.
In November the company announced that it had moved its three dimensional memory to a 0.15-micron manufacturing process technology.
A third company working in the third dimension is Tezzaron Semiconductor Inc. (Naperville, Illnois), which bonds devices together at the wafer-level, to stack a microcontroller chip with an SRAM memory. By designing the two chips for this configuration the distance between memory and ALU can be significantly reduced and performance increased as a result, the company asserts.
In November a partnership with MagnaChip Semiconductor Ltd., the renamed former non-memory operations of Hynix Semiconductor Inc.
Initially, MagnaChip built a RAM-based register file on a foundry basis for Tezzaron. "It's a relatively primitive device," according to Robert Patti, chief technology officer for Tezzaron. "But we believe that this will usher in a new era of semiconductors." Tezzaron itself plans to develop and sell a range of 3D "custom devices," Patti said. 3D chips hold great promise in terms of speed, density, and low power requirements for cellular phones, handheld devices and other products.
(Return to the 2004 Top 10 story list.)
|
|
|
|
|
