The concept of 3D DRAM is being discussed, and much work is being done in the industry to create the machinery, advanced ALD, selective deposition, selective etch, and 2D materials for access devices. The simplest method for stacking DRAM cells is placing several dies on top of one another while maintaining the status quo in terms of DRAM technology.
In contrast to the current architecture, there are new techniques for creating dynamic memory cells that are more suitable for monolithic stacking. The majority of the work being done now, however, is being put into stacking numerous DRAM dies, which enables the current memory cell, which has had years to learn, to maintain its supremacy.
The Global 3D DRAM market accounted for $XX Billion in 2021 and is anticipated to reach $XX Billion by 2030, registering a CAGR of XX% from 2022 to 2030.
Samsung Electronics are accelerating the R&D of 3D DRAMs. The industry titan in semiconductors has begun to support allied groups by hiring staff. Drams were once created by placing transistors and capacitors in a straight line. However, once DRAM capacity surpassed 4 megabits, it became challenging to boost DRAMs’ density, necessitating circuit and capacitor reorganisation.
The DRAM market at the time was split between “trench” manufacturers, who elected to bury circuits and storage devices beneath planes, and “stack” manufacturers, who opted to stack them on planes. The CAA configuration transistor 3D DRAM, which is based on the indium gallium zinc IGZO-FET material and has outstanding temperature solidity and dependability, has been made available by Huawei.
NEO Semiconductor, a pioneer in creating innovative architectures for 3D NAND flash and DRAM memory, announced the introduction of X-DRAM, its newest technological advancement. By reducing power consumption and boosting the performance of the main memory used in IT systems and consumer products, X-DRAM offers remarkable gains over the standard DRAM.
A vertical channel-all-around (CAA) transistor has been proposed by Huawei, a leading manufacturer of communications equipment in China, which may be appropriate for the creation of 3D-DRAM.
The component is a vertical column-shaped indium gallium zinc oxide (IGZO) field effect transistor (FET) made up of layers of IGZO, high-k dielectric hafnium oxide, and IZO. The thickness of the IGZO is about 3nm. About 8 nm thick, the HfOx and IZO. The critical dimension in-plane is 50nm, and the channel length in the vertical direction is 55nm.
The transistor achieves a sub-threshold swing of 92 mV/decade and a current density of 32.8 microamps/micron at Vth plus 1V. The transistor is a promising candidate for high-performance 3D DRAM beyond 1-alpha nodes in the future, according to the authors, who claim good thermal stability and reliability from -40 degrees C to +120 degrees C.
These thoroughly researched technologies are creatively combined by Monolithic’s 3D DRAM technology: monolithic 3D with shared litho steps between several memory layers, Ion-cut, double gate, single crystal Si, floating body RAM cell with body-stored charge. Leading DRAM manufacturers and major equipment suppliers are already considering monolithic 3D DRAM, the DRAM equivalent of 3D NAND, as a potential solution for long-term scaling.
Thanks to research from IBM and Micron, smaller, faster RAM with up to 128GBps transfer speeds might soon be available. The companies have created three-dimensional memory by vertically stacking separate DRAM chips that would typically need to be placed side by side.
However, the efficiencies gained aren’t just due to saving on space. Through-silicon vias, or TSVs, a novel invention that runs vertically through the stack of chips and serves as a conduit to the host device, enable communication between the stacked chips and the device.
The TSVs allowed the memory to test at 128GBps, which is ten times faster than the current memory. Additionally, IBM asserts that the chips are 70% more energy-efficient than current DRAM. The Hybrid Memory Cube technology developed by Samsung and Micron is anticipated to heavily rely on the new research. In two years, IBM and Micron expect it to be commercially available, with servers probably using it first.
© Copyright 2017-2023. Mobility Foresights. All Rights Reserved.