Cover Story
a choice of offering liquid or solid. If the process is developed, then the POR (Process of Record) is set. Taking all of these developments into
16
account, there are significant opportunities for companies that understand these new material operating environments and can provide these materials, specialised chemistries and processes that underpin the ongoing evolution of the semiconductor market. Making new memories With the emergence of new memory iterations there are accompanying opportunities for materials suppliers. Ultimately, of course, the success in the marketplace of memory materials will be tied to the types of memory that are in production and dominant in end user devices but, as the drive for increased device performance continues and the pace of materials adoption and precursor development intensifies, the resulting compacted timelines are placing added emphasis on collaborative research and development in the memory space. The use of high-k layers has enabled the continued shrinkage of CMOS transistors and DRAM capacitors, allowing a larger number of devices per unit area and, in the case of CMOS, the continued following of Moore’s law. However, limitations lie ahead very similar to those experienced when silicon-based materials started to reach their threshold. While these high-k materials are currently continuing to move forwards to replace older technologies, in all likelihood they themselves will not be used in their existing forms in, say, 15 to 20 years’ time. As we move beyond the 32nm node, the
Figure 2: Memory Technology Classifications
(courtesy of An Chen, Global Foundries)
circuitry in a typical device becomes finer, and at the point DRAM, SRAM and NAND Flash will begin to reach their physical limitations in terms of their continued scaling, becoming increasingly expensive and difficult to manufacture. Critical
then is the future integration of materials at the gate level and new material requirements of competing non-volatile memory technologies which are emerging to overcome the scaling limits of current memory technologies. These include PCM and a host of emerging memory iterations being evaluated by collectives such as the ITRS’ Emerging Research Devices (ERD) and the Emerging Research Materials (ERM) Working Groups. Figure 2 illustrates the classification of memory technologies, where the prototypical classes are either in play or fairly imminent, with emerging areas being ‘ones to watch.’ DRAM is not going to become obsolete overnight and, in the short-term at least, its market will remain significant and may even increase as the 32nm node rolls out to mass markets. However, having one eye on the future, as the semiconductor industry invariably does, and looking at the challenges facing DRAM beyond 32nm, there is an onus on manufacturers and materials suppliers to deliver on emerging memory options, processes and product categories through ongoing collaborative Research and Development efforts. Speaking from a material supplier’s perspective, SAFC Hitech has already developed an in-depth understanding of phase-change memory (PCM) and offers a portfolio of molecules developed in-house. As we see these materials moving into the mainstream with customers assessing their feasibility for thin film fabrication processes, we are constantly looking to evaluate what’s next in memory. Currently, SAFC Hitech’s high k materials are used in high capacitance dielectric layers for DRAM, CMOS, eDRAM and Flash architectures, and its precursors are increasingly being used in additional functional memory layers such as MRAM, ReRAM, FeRAM, and PCM memory architectures.
One of the main drivers of the development of new types of memory is mobile computing and data, which has seen the emergence of ‘cloud’ based technologies, with core programs not resident on devices but existing instead in vast data farms that can only be accessed via applications through user interfaces. Conventional DRAM memory technologies, which requires the re-writing of data several times per second, and the devices that house them, such as traditional spinning hard drives, are limited for such high volume cloud-based applications, as they consume a great deal of power and generate significant amounts of heat. A number of technologies are potential
www.euroasiasemiconductor.com Issue IV 2011
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40