Introduction to FRAM (Ferroelectric RAM) including Its History [MiniTool Wiki]
Introduction to FRAM
Ferroelectric RAM (FRAM, FeRAM, or F-RAM) is a random-access memory that combines the fast read and write access of DRAM (Dynamic RAM), while also using a ferroelectric layer instead of a dielectric layer to achieve non-volatility. FRAM is one of the alternative non-volatile random-access memory technologies that provide the same functions as flash memory.
Although the name FRAM or ferroelectric RAM seems to indicate the presence of iron in the memory, it is not. The data retention time of FRAM at +85°C exceeds 10 years (up to decades at lower temperatures). And the market disadvantage of FeRAM is that the storage density is much lower than that of flash memory devices, the storage capacity is limited and the cost is higher. Like DRAM, FeRAM's read process is destructive, so it requires a write-after-read architecture.
- In 1952, Dudley Allen Buck proposed ferroelectric RAM in his master’s thesis - Ferroelectrics for Digital Information Storage and Switching.
- In 1955, Bell Telephone Laboratories began to study ferroelectric-crystal memories.
- After the introduction of metal-oxide-semiconductor (MOS) dynamic random-access memory (DRAM) chips in the early 1970s, the development of FRAM began in the late 1980s
- In 1996, Samsung Electronics introduced a 4 Mb FRAM chip manufactured using NMOS logic circuits.
- In 1998, Hyundai Electronics (now SK Hynix) also commercialized FRAM technology.
- In 2000, Sony released the first commercial product using FeRAM - PlayStation 2 (PS2).
- In 2001, Texas Instruments (TI) cooperated with Ramtron to develop FRAM test chips with an improved 130 nm process.
- In 2005, Fujitsu and Seiko-Epson jointly developed the 180 nm FRAM process.
FRAM occupies only a small part of the entire semiconductor market. In 2005, global semiconductor sales were $235 billion (according to the Gartner Group), of which the flash memory market accounted for $18.6 billion (according to IC Insights).
According to reports, Ramtron, which may be the largest FRAM supplier, had sales of $32.7 million in 2005. Compared with alternative NVRAM, flash memory sales are much larger to support greater R&D efforts.
Samsung (2007) uses a semiconductor linewidth of 30 nm to produce flash memory, while Fujitsu (Fujitsu) uses a linewidth of 350 nm for production, and Texas Instruments (2007) uses a linewidth of 130 nm to produce FeRAM.
The flash memory cell can store multiple bits in each cell (currently 3 bits in the highest density NAND flash memory device), and due to the innovation of flash cell design, the number of bits per flash cell is expected to increase to 4 or even to 8. As a result, the area density of flash memory is much higher than that of FRAM, so the cost per bit of flash memory is several orders of magnitude lower than that of FRAM.
In 2005, Ramtron reported a large number of sales of its FeRAM products in various fields, including (but not limited to) electricity meters, automotive (such as black boxes, smart airbags), business machines (such as printers, RAID disk controllers), and instrumentation, medical equipment, industrial microcontrollers, and radio frequency identification tags. Other emerging NVRAMs, such as MRAM, may try to compete with FeRAM to enter a similar niche market.
Texas Instruments (TI) demonstrated that in the traditional CMOS semiconductor manufacturing process, two additional masking steps can be used to embed FeRAM cells. Flash usually requires nine masks. For example, this can integrate FeRAM into a microcontroller, and the simplified process can reduce costs.
However, the materials used to manufacture FeRAM are not commonly used in CMOS integrated circuit manufacturing. Both the PZT ferroelectric layer and the noble metal used for the electrodes increase the compatibility and contamination problems of the CMOS process. Texas Instruments (TI) has integrated a certain amount of FRAM memory in its new FRAM series of MSP430 microcontrollers.
Pros and Cons
Compared with flash memory, the advantages of FRAM
- Lower power usage
- The larger number of write-erase cycles
- Faster write performance
Compared with flash memory, the disadvantages of FRAM
- Lower storage density
- Higher cost
- Overall capacity limitation
FRAM can provide many advantages and can be used in many fields, but in many cases, the use of FRAM memory is a balance between many characteristics and parameters, and these characteristics and parameters need to be designed for any specific circuit design.
To sum up, this post gives you some detailed information about FRAM (ferroelectric RAM), including its definition, history, usage as well as its pros and cons.