Ann: Q2FY19 Activities Update and Appendix 4C, page-83

Thanks Khoutam for raising the theoretical advantages of the two technologies. I use the term theoretical because at the present time the potential advantages of non-filamentary reram read well on the slide you keep using but there is more to it than that.

Reram relies on complex effects to work. We still dont know if the 4DS design actually works. Kind of like how we all know how an airbag works but that didnt stop the engineers at Takata stuffing it up.

There is lot more than the filamentary nature of the design in getting working, perfomant chips. 4DS first wafer didnt work. Obviously nothing to do with the non filamentary nature of the tech. Perhaps the metal layering was not right or separation effects occured.

There are many factors beyond the cell that impact scalability. In theory non filamentary cells should scale better. But what about the rest of the supporting logic? Perhaps the care the Leti folks are taking with assessing scalability means more design care to avoid such effects.

I have watched the use of "exotics" like Gallium Arsenide over the years and they invariably cost more than the current prevailing standard. That is fine but usually yields are lower and prod processes more complex and costly. What material is 4DS again? Perovskite. Its getting a lot of research attention of course but its properties for reram are still being explored.

Refer to Article in Journal of Materials Chemistry C 5(23) · April 2017 

To probe the performance of perovskite memory devices: defects property and hysteresis Ziqi Xu,a Zonghao Liu,b Yuan Huang,a Guanhaojie Zheng,b Qi Chenc and Huanping Zhou*ab Author affiliations Abstract Hybrid organic–inorganic perovskite materials offer a range of interesting characteristics that are suitable for optoelectronic devices, such as photovoltaics. Along with the fast rise in device performance, a current density–voltage (J–V) hysteresis originating from defects and their movement has attracted intense attention, which renders challenges regarding the stability and reliability of the novel materials. Here, we carefully probe the effects of defects in perovskite materials and across interfaces within the device, in which bistable conductive states are achieved for the next generation of nonvolatile memory. The memory device shows an operating voltage as low as 0.25 V, and a decent ON/OFF ratio. More importantly, we correlate the defect density and hysteresis-index of different perovskite films with the corresponding memory device performance. The findings enrich our understanding of the working mechanism of perovskite memory devices, which will also benefit other organic–inorganic hybrid perovskite optoelectronics.

The 4DS reram story is compelling, and I am watching closely. However there is one thing about solving hard problems we should all remember and that is, that it's hard.