ReRAM Gains Momentum in ‘Universal Memory’ Race
By Mark LaPedus, SemiMD senior editor
Resistive RAMs (ReRAMs) are one of several next-generation candidates to succeed NAND flash or other memory types, but there are material, production and cost issues associated with the technology.
Elpida, Hynix, IMEC, Micron, Panasonic, Samsung, Sharp and others are working on ReRAM. In fact, Micron and Sony Corp. have quietly forged an alliance in ReRAM. “Micron and Sony have entered into a joint development program (JDP) to co-develop a new non-volatile memory focused on a conductive bridge-type random access memory, ReRAM,” according to a spokesman for Micron.
“Micron entered into the JDP with Sony as part of the company’s regular research and development efforts that involve, among other work, researching various emerging memory technologies,” according to the spokesman.
And this week, IMEC, Macronix, Panasonic and others presented ReRAM papers at IEDM. ReRAM is one of several “universal memory” technologies. Phase-change, MRAM, FRAM and others fall into that category. They are trying to replace DRAM, NAND, NOR or all three. But the problem is the next-generation memory types are still not in production despite years of R&D. They are expensive to make and difficult to scale.
ReRAM is a product that holds potential to replace NAND. “Current charge storage based flash memory technologies are believed to face scaling limitations beyond 18nm,” according to IMEC. “To overcome these, a variety of innovative cell and memory concepts are investigated worldwide. One of the most promising memory concepts is the resistive RAM or RRAM. It is based on the electronic switching of a resistor element material between two stable (low/high) resistive states. The major strengths of RRAM technology are its potential density and speed.”
There are challenges. “However, even if many materials reported to date exhibit good resistive switching properties, the success of a future RRAM technology depends critically on integrability into a conventional, underlying baseline technology, with cost as a key factor,” according to IMEC.
Top-view SEM picture of IMEC's processed ReRAM cell (Source: IMEC)
The cell has fast nanosecond-range on/off switching times at low-voltages. It has a large resistive window (>50) and shows no closure of the on/off window after functioning at 200°C for 30 hours. The device even remained operating failure-free functioning for 30 hours with a thermal stress of 250°C. The switching energy per bit is below 0.1pJ, and AC operating voltages are well below 3V. With these characteristics, IMEC’s cell meets the major requirements for device-level nonvolatile memory.
These results were obtained in cooperation with IMEC’s key partners in its core CMOS programs Globalfoundries, Intel, Micron, Panasonic, Samsung, TSMC, Elpida, Hynix, Fujitsu and Sony.
Another IEDM paper was given by the Nanyang Technological University, Institute of Microelectronics, Peking University, A*STAR, National University of Singapore, GlobalFoundries, Soitec and Fudan University: “We report a high performance, forming-free and self-rectifying unipolar HfOx based RRAM fabricated by fab-available materials. Highlight of the demonstrated RRAM include 1) CMOS technology friendly materials and process, 2) excellent self-rectifying behavior in LRS (>103 @ 1 V), 3) forming-free unipolar resistive switching, 4) wide read-out margin for high density cross-point memory devices (number of word-line >106 for worst case condition).”
The big memory houses are also exploring ReRAM. Last year, for example, Hynix Semiconductor Inc. entered into a joint development agreement with HP to develop memristor technology in the form of ReRAM. The two companies will jointly develop new materials and process integration to deliver ReRAM to market by transferring the memristor technology from research to commercial development. Hynix will implement the technology in its R&D fab.
The memristor, short for “memory resistor,” requires less energy to operate, can retain information even when power is off, and is faster than present solid-state storage technologies.
Meanwhile, Micron, along with Sony, are exploring ReRAM. The move into ReRAM represents Micron’s latest effort in next-generation memories. The company entered the phase-change memory race, when it recently acquired Numonyx, a supplier of NOR flash devices.
Numoynx, formerly the NOR flash units of Intel and STMicroelectronics, was developing phase-change memory. Samsung Electronics Co. Ltd. is also separately developing phase-change memory.
Earlier this year, Unity Semiconductor Corp. entered into a joint development agreement with Micron. Micron invested in Unity. For the last eight years, Unity has been developing CMOx, a next-generation memory type. CMOx is designed to scale beyond the limitations of the legacy transistor technology currently used in NAND flash memory.
More recently, Micron rolled out the Hybrid Memory Cube (HMC), a 3D DRAM technology for servers and networking systems. In November, Micron and Singapore’s A*STAR Data Storage Institute (DSI) jointly announced that the two companies entered into an agreement to collaborate on the development of spin transfer torque magnetic random access memory (STT-MRAM), a promising alternative non-volatile memory technology for next-generation storage.
As part of the collaboration, Micron and DSI will invest in joint research to develop high-density STT-MRAM devices during the next three years. Meanwhile, Samsung and Hynix are developing STT-MRAM. In fact, Samsung recently bought Grandis, a developer of STT-MRAM. Toshiba and others are also exploring the technology.
Micron is also looking at all next-generation memories. “Considering Micron’s previous high visibility acquisition of Intel’s phase change memory (PCM) program, these new announcement are likely part of the regular activity of all memory companies to keep abreast of any potentially critical new technologies,” said Bob Merritt, an analyst with Convergent Semicondctors LLC, in a blog.
“However, Micron already has a long history in ReRAM that can be traced back to their 2002 licensing of Axon Technologies’ Programmable Metallization Cell (PMC) technology,” he said.
“While the public perception of memory technologies tends to assume that no new technology will be acceptable until it reaches the same cost per bit of DRAM or NAND, we believe that the continued interest in these new and emerging technologies is based on finding other market entry points and an expectation of providing high value to new applications,” he added.
Axon Technologies Corp. has also been trying to commercialize PMC. Adesto Technologies, which is backed by Applied Materials and others, is developing conductive bridge RAM (CBRAM) memory technology. CBRAM was originally developed at Arizona State University and is also known as PMC.
