Global Sources
EE Times-Asia
Stay in touch with EE Times Asia
?
EE Times-Asia > Optoelectronics/Displays
?
?
Optoelectronics/Displays??

Addressing PCM elevated temp data retention

Posted: 01 Sep 2015 ?? ?Print Version ?Bookmark and Share

Keywords:phase change memory? PCM? Ge? segregation model? Elevated temperature data retention?

Figure 2 illustrates the benefits of this for the case where low current SET pulses are used, both with and without a trailing edge. The result of forming is in the crystallized GST that acts as the primary part of the upper electrode a region of crystallized material of fixed orientation is created as illustrated in figure 2(a). This then acts as the permanent template, or seed, for a similarly oriented conducting column of the low resistance SET state as it grows from the cylindrical surface, figure 2(b). It is not clear in the growth of this column acting as a seed is responsible for the Ge segregation, a sort of pseudo-epitaxial zone refining, or some other mechanism is responsible.

Figure 2(c) illustrates the randomly oriented polycrystalline structure after a high current SET. The inter-grain amorphous material accounts for the lower stability of the SET state with respect to resistance drift.

Figure 2: (a) A RESET device illustrating the single oriented crystal template or seed formed outside of the active volume. (b) A low SET current with or without a trailing edge results in a conducting filament with a single crystal orientation, (c) The randomly oriented micro-crystallites resulting from high current SET with inter crystal amorphous material.

There are then two elements to the forming process, one to complete the Ge segregation and the other to create the singly oriented template beyond the dome or half-cylinder surface.


Pros and cons
The results summarise in the preceding paragraphs were for cells of the type shown in figure 3 from a 12Mb test vehicle on a 90nm process, using the edge type PCM cell structure. W/E endurance up to 10^7 cycles with 30ns RESET and 800ns SET pulses was demonstrated, with an on-to-of ratio of greater than two orders of magnitude maintained throughout the test. The radius of the cylindrical part of the active region was ~50nm.

Figure 3: Sub-lithographic edge-type PCM cell illustrating the normal cylindrical active volume (purple).The heater electrode (orange) would be expected to have dimensions of ~10nm x 70nm for an aggressive 90nm node lithographic process.

The proof of high temperature performance had a number of elements, one of which was baking ~80 cells of the Ge rich alloy in temperature steps up to 240C in one-hour increments. This showed a wide on-to-off resistance ratio could be retained, with what appeared to be annealing of the RESET resistance state for an increase of about two orders of magnitude. The SET state remained constant, it is suggested that the single crystal orientation of the SET column provides stability of the SET state and removes drift.

The value of the RESET resistance with annealing for one hour at temperatures up to 240C for devices after 10 and also 10^7 W/E cycles was measured. The starting point showed a difference in RESET resistance of about an order of magnitude with the devices subjected to 10^7 cycles having the lower resistance (i.e 10^6?). Again, the resistance increased with anneal temperature until the crystallisation temperature.

If the lateral segregation and forming takes place over the initial few W/E cycles then the results of baking for cells in the 1-10 W/E cycle range would be of more interest. Although the authors of the paper did demonstrate the availability of a wide working window with variations of 2at% Germanium in composition.

While the Ge segregation model appears to work for the edge half-barrel cylindrical or dome amorphous structure, as scaling necessitates a move to the pore structure it will not be possible to have a conducting cylindrical layer of crystallized Ge surrounding the optimised active area. In fact in a PCM "pore" structure where all of the material in the confined in the pore will be switched, the required segregation in the radial direction may not be possible. Does that mean the problem of element separation or segregation will not occur? If it does the options are localized regions of concentration or composition changes in the direction of current flow towards one electrode. In PCM pore structures the radial segregation of crystallized Ge may well account for the failure to RESET reliability problems.


Reference
[1] Operation Fundamentals in 12Mb Phase Change Memory Based on Innovative Ge-Rich GST Materials Featuring High Reliability Performance, by V Sousa et al, CEA-LETI and STMicroelectronics, Proc VLSI 2015.

About the author
Ron Neale is an independent electrical/electronic manufacturing professional.


?First Page?Previous Page 1???2



Article Comments - Addressing PCM elevated temp data re...
Comments:??
*? You can enter [0] more charecters.
*Verify code:
?
?
Webinars

Seminars

Visit Asia Webinars to learn about the latest in technology and get practical design tips.

?
?
Back to Top