Global Sources
EE Times-Asia
Stay in touch with EE Times Asia
EE Times-Asia > Memory/Storage

Recent revelations on non-volatile memory conduction

Posted: 24 Dec 2015 ?? ?Print Version ?Bookmark and Share

Keywords:conduction? non-volatile memory? phase change memory? PCM? GeTe?

The ongoing project [reference 1] at IBM Zurich to meticulously detail every aspect of the conduction processes involved in non-volatile memory materials and devices, phase change and oxides, has just provided us with some new and important insights, as well as an intriguing mystery.

If any of the candidate NV memory technologies, chalcogenide glass or oxide based, are ever going to succeed in replacing silicon and meeting future mass memory requirements, that success will only be based on detailed understanding. The lessons of the past 50 years with phase change memory (PCM) and what can only be described as commercial, try-it-and-see, disasters make that point abundantly clear.

The latest work [reference 1] explores conduction in the pre switching sub-threshold region [reference 1] and the results promised for a later date will concentrate on the next part of the I-V curve threshold switching and beyond.

Among the many theories for electrical conduction in PCM materials prior to the point where threshold switching occurs some version of Poole-Frenkel conduction is most often cited. In the past three separate regions of electrical conductivity have been clearly recognised. They are an Ohmic region, followed a region with low to moderate applied electric fields, described as Poole conduction and finally a high field region of Poole-Frenkel conduction. With the latter two involving the release of carriers by lowering of the Coulomb barriers by the effect of electric field on near neighbour traps.

While it is possible to account for each region individual theoretical descriptions of the different conduction process only prove to be applicable over a limited range of the observed characteristics. The IBM, Zurich team have developed a new model with which they are now able to link together in a continuum the three recognised regions of sub-threshold switching conduction.

The strength of the new model can be seen in figure 1 where the the generally accepted curves for Poole (green) and Poole-Frenkel (orange) conduction from the literature have been overlaid on the measured and those calculated from the new model. In the example of figure 1 a dome structure PCM device was used.

Figure 1: Comparison of original experimental results overlaid with those from the new model for nanoscale GeTe PCM cells with Poole (green) and Poole-Frenkel (orange) conduction.

The measured and results from the model for a GeTe material composition are coincident. The model has also been verified for large area thin films, gap structures and nanometric sized PCM sized device structures of both recent and historical origin, including the effects of temperature and drift. It is considered the new model will have general applicability to the amorphous oxides used in ReRAM/RRAM as well as PCM materials. In the new model there is now a single equation with which it is possible to describe electrical conductivity over all of the sub-threshold switching regions, it is highly complex and includes an integral.

The table lists the different PCM device structures and films used to demonstrate the accuracy of the model with the constants used in the model, some measured, others from room temperature curve fitting.

Table: Important parameters and the variety of experimental devices and compositions used to evaluate the new IBM model. The estimated nearest neighbour trap distance (s). Ea the Ohmic region activation energy, εr the relative permittivity, uo the zero field mobility and K a constant. K, u0 and s are obtained from room temperature curve fitting. (Measured green box, estimated dark blue box, literature sourced light blue box.)

Velocity saturation
One of the keys to the new model, and possibly the major one, is the introduction of a new constant called velocity saturation. A tempting although not strictly accurate analogy is to liken saturation velocity to the terminal velocity of an object moving through through a fluid, say a free fall parachutist. The more serious investigation of velocity saturation has a starting point which suggests it is caused by optical phonon scattering as is the case in crystalline material. The value attributed to this constant in the new model is 50V/?m.

1???2???3?Next Page?Last Page

Article Comments - Recent revelations on non-volatile m...
*? You can enter [0] more charecters.
*Verify code:


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

Back to Top