Fujitsu Microelectronics America announced that the Tokyo Institute of Technology, Fujitsu Laboratories, and Fujitsu Limited have jointly developed a new material for a new generation of non-volatile Ferroelectric Random Access Memory (FeRAM).
FeRAM combines the fast operating characteristics of DRAM and SRAM with flash memory's ability to retain data while powered off. FeRAM, like MRAM, is a low power, nonvolatile memory design that possesses an array structure allowing for the one-transitor memory
cell density of standard DRAM. While MRAM derives its memory store capability from the
presence of a magnetic junction tunnel, FeRAM utilizes the properties of ferroelectric
capacitors to effectively detect the positioning of electrical dipoles within
ferroelectric material. Unlike DRAM memories, which refresh their capacitors
periodically, ferroelectric memories only refresh the arrays after a read is performed,
this because the read destroys the content of the cell. If there is no read, the dipoles
never flip, hence no recharge cycles are required. This allows FeRAM memory to operate
at low power.
The material is a modified composition of Bismuth Ferrite (BiFeO3 or BFO), which enables
data storage capacity up to five times greater than the materials currently used in
New FeRAMs can be produced with Fujitsu's 65nm process technology using the BFO-based
material in a device structure similar to the one used to build FeRAMs using 180nm
technology. FeRAMs using this material can provide memory cell capacity up to 256Mbits.
The new FeRAMs will deliver the very low power consumption and high speeds required for
new generations of personalized mobile electronic products such as IC cards, which must
be small, easy to use, and provide very high security. Engineering sample shipments are
planned for 2009.
Fujitsu began mass production of FeRAM in 1999, and has shipped several hundred million
chips of embedded and stand-alone FeRAM as of March 2006, including 1Mbit stand-alone
FeRAMs built using 65nm technology can be produced using Mn-doped BFO, with the similar
device structure of the FeRAM being produced using 180nm technology. FeRAMs using this
new material will also provide significant scalability, enabling large memory capacity
up to 2014.
With further development of BFO, large-capacity 256Mbit FeRAMs can be realized that
reach densities of two orders higher compared with the current capacity of 1Mbits. With
this increased density, FeRAM applications is expected to expand not only in the
security applications, but also in new domains like "Quick-On-Computer," under which a
computer can be immediately ready to use after turning on. FeRAMs also can be used in
electronic paper devices, which let users browse and read a large volume of information
traditionally written on plain paper.
Details of the material and process were presented at the meeting of the Japan Society
of Applied Physics (JSPS) in March 2006 in Tokyo, and also at the International
Symposium on Integrated Ferroelectrics (ISIF) in April 2006 in Hawaii. For the joint
research, Tokyo-Tech is supported in part by a grant from the Ministry of Education,
Culture, Sports, Science and Technology (MEXT) via the Japan Science and Technology
Fujitsu will continue its research and development programs to provide FeRAMs for
embedded LSI. Research on the BFO material and implementation technology needed for
incorporating BFO capacitors in FeRAM also continues.
BFO is a ferroelectric material composed of Bismuth, Iron and Oxygen atoms with a
perovskite structure. Lead Zirconate Titanate (PZT or Pb(Zr,Ti)O3) is now used as a
ferroelectric material but it has a lower-charge storage capability, so it also has
limited scalability. The technology limits of PZT are expected to occur at the 130nm
node, because as cell area decreases, higher polarization is required. This limit is
expected to be reached in 2009.
An Mn-doped BFO thin-film capacitor was developed with the dual functions of decreasing
leakage current and 180-220 C/cm2 of switching charge, Qsw, which is equivalent to
twice the remanent polarization, 2Pr. These results clearly indicate significant
scalability potential for future technology nodes.