A significant amount of research and development (R&D) is targeting the next generation of electronic devices as there is an increase in the manufacturing of smart fixed, portable, and wearable electronic devices. During the past 5 years, the proliferation of smartphone and wearable devices has created a huge demand for the memory market. At the same time, the memory storage used in these devices is growing exponentially (doubling every 2 years), and users are desperate for a higher-performance memory device. Electronic device performance is based on non-volatile memory, and the most widely used form of primary non-volatile memory to increase system performance is random access memory (RAM). As the size of electronic devices shrinks, components placed inside these devices should inevitably have a proportionally lower form factor, without compromising the efficiency and performance of the device. Scaling RAM, therefore, is a complex process and a challenge for memory manufacturers. To date, flash RAM has dominated the entire non-volatile memory market because of its high performance and scalability. However, as flash RAM has reached its scalability limits, the emerging RAM technology has substantial competition, which can replace or disrupt the non-volatile memory market.
Limitations of the existing memory technology
The main challenge in the non-volatile memory market is to reduce scalability and increase performance as wireless wearable devices, smartphones, and Internet of Things (IoT) devices need more memory with less energy and higher performance. Semiconductor development is based on scaling because of the shrinking size of smart devices. With increased use of wearable devices and cloud computing, more storage capacity is needed, and RAM’s storage capacity is increasing by 90% each year. In addition, with intense developments in electronic devices, the requirement for performance, speed, capacity, and lower energy has increased. Hence, to meet the form factor requirement of the electronic device manufacturers with increased RAM capacity and performance, transistors placed in the flash RAM are doubled. This results in reaching the maximum scalability limit of 16 nanometers for flash RAMs. Because of scalability limits in the existing memory technology, the device manufacturers will move from flash transistors to emerging memory.
Emerging ReRAM memory technology disrupting the existing memory technology
Emerging memory technologies have been present in the market for the last decade, but have not gained momentum because of the dominance of flash RAMs. The market mainly has 3 types of emerging memory technologies that are capable of replacing flash RAMs: ReRAM, phase-change memory (PCM), and magnet RAM, out of which the most effective is the ReRAM. The ReRAM is a niche market in 2016 and will be a mainstream of memory by 2020. Many companies are developing innovative ReRAM technology solutions focused on the future memory market. This quantum leap will allow semiconductor memory elements to become cheaper, faster, more reliable, and more energy efficient than existing flash RAM technology.
ReRAM’s main advantage is scalability because the existing flash RAMs consists of transistors that have 3 terminals: source, drain, and gate, but the ReRAM will have a resistor with only 2 terminals: anode and cathode. This main advantage allows for a denser number of resistors in the same size, comparatively. The second advantage is that flash memory is unable to scale below the limit of 16 nanometers because the flash cells will confine electrons; therefore, electrons cannot be confined if scaling lower than its limits. In the ReRAM technology, however, confining the electrons is not necessary as the memory is programmed by changing the material properties and composition, thus the resistivity is changed. This factor allows the ReRAM technology to be scaled below 16 nanometers. Currently, the highest capacity available in flash RAM is 256 gigabytes, and the lowest scalability achieved by flash RAM is 16 to 20 nanometers. Weebit, a ReRAM manufacturer, has achieved a scalability of 5 nanometers and can increase its memory capacity up to 1TB.
The third advantage is that to program the flash cells, a high voltage of about 20 volts (V) is needed (the available power in a smartphone is 5 V); therefore, internal charge pumps are used to provide 20 V. This technology increases the voltage internally from 3 V to 20 V, which will require space for the charge pumps and capacitors to store the charge and consume more energy. With ReRAM, only 2 to 3 V are needed without using charge pumps, resulting in lower energy and faster speeds. To program flash cells, the existing data must be erased; therefore, the time taken to erase and program will lower the response time of the flash memory, which is not the case in ReRAM. Thus, ReRAM has the competitive edge over the flash memory technology in terms of improved performance, low energy, and scalability.
Conclusion
The emerging memory market has been developing for the past 10 years but has not reached any application because flash RAMs have had a lot of momentum in the non-volatile memory market and will keep on scaling until it reaches the scalability limit. To date, flash RAMs are more efficient, scalable, and fast, matching the current technology advancements. In the next 5 years, however, with the scalability limits of the flash RAM, the need for emerging memory technology will replace flash RAMs. The ReRAM has a simpler structure and requires minimum space. In addition, among emerging memory technologies, the ReRAM is more cost effective and efficient as it requires few masks in the silicon bed and can be integrated in a MCU.
