Engineering of chalcogenide materials for embedded applications of Phase Change Memory

Zuliani, P.; Palumbo, E.; Borghi, Massimo; Libera, G. Dalla; Annunziata, R.

doi:10.1016/j.sse.2015.04.009

Cited by 60 publications

(39 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Retention instability at high temperature has been the main challenge for PCM technology in the last decade, in particular to target Industrial and Automotive applications. Exploration and engineering of new alloys showed that PCM can retain the information up to temperatures higher than 200 • C. STMicroelectronics demonstrated the capability of a GeSbTe based material called "T-alloy" to grant a retention of one hour at 240 • C without data loss [6]. Being this result compatible with automotive specifications (~2 years at 150 • C) and soldering reflow thermal profile (peak temperature equal to 260 • C, according to JEDEC standard), it opens the way for PCM towards the embedded market.…”

Section: Pcm High Temperature Data Retention and Programming Speedmentioning

confidence: 99%

Phase-Change Memory: Performance, Roles and Challenges

Navarro

Bourgeois

Kluge

et al. 2018

2018 IEEE International Memory Workshop (IMW)

View full text Add to dashboard Cite

In this paper we analyze recent progress in Phase-Change Memory (PCM) technology targeting both Storage Class Memory and embedded applications. The challenge to achieve a high temperature data retention without compromising the device programming speed can be addressed by material engineering. We show that volume and thermal confinement improvement of the phase-change material enables a high (10-fold) reduction of the programming current, achieved also by the optimization of the device architecture, in particular in the case of a confined structure. It leads to a higher cell efficiency proven by a 6x reduction of the programming current density wrt a standard PCM structure. Furthermore, we demonstrate the reduction of thermal losses by the tuning of the thermal conductivity of the dielectrics surrounding the phase-change material. Finally, we propose some considerations about the PCM ultimate scaling and the reliability at such dimensions, showing that the engineering of the bottom electrode/phase-change material interface can lead to a reduced variability in scaled devices.

show abstract

Section: Pcm High Temperature Data Retention and Programming Speedmentioning

confidence: 99%

Phase-Change Memory: Performance, Roles and Challenges

Navarro

Bourgeois

Kluge

et al. 2018

2018 IEEE International Memory Workshop (IMW)

View full text Add to dashboard Cite

show abstract

“…Generally speaking as represented in Figure 7(a) phasechange materials from Ge-Sb-Te ternary diagram are known to feature excellent performances in data retention if moving toward Ge-rich compounds [11], while highest programming speed is achieved if moving toward Sb-rich compounds [12]. For instance, in the Ge-Sb-Te ternary diagram, it is well known that going toward Sb-rich GST, moving on the GeTe-Sb pseudo-binary tie line on Ge1SbxTe1 compounds (x≥1), provides materials that feature faster and faster crystallization speed, and correspondingly the crystallization temperature of amorphous layers goes lower, as represented in Figure 7(b) [12].…”

Section: Pcmmentioning

confidence: 99%

“…[13]- [14], exhibit much higher crystallization temperature and slower crystallization speed at material level; implying at device level better data retention (Figure 8(a)) and lower SET speed (Figure 8(b)), respectively. Such class of material is indeed most suited for embedded applications (ie automotive, smartcard) [11].…”

Section: Pcmmentioning

confidence: 99%

Universal Signatures from Non-Universal Memories: Clues for the Future...

Perniola

Molas

Navarro

et al. 2016

2016 IEEE 8th International Memory Workshop (IMW)

View full text Add to dashboard Cite

The quest for the universal memory has been pursued since several years, but as far results from scientific literature do not declare one single technology able to fit all the requirements of the memory hierarchy. Flash and DRAM cover more than 95% of the global sales in the semiconductor memory market [1], but none of the two is able to substitute the other. This appears to be true also for disruptive backend memory technologies, like OxRAM, CBRAM, PCM and MRAM. This paper deals with universal signatures that stand true for such disruptive technologies. A specific analysis on two case examples from the RRAM and the PCM domain, where tradeoffs can be adjusted to target specific applications, is finally proposed. Knowing in advance the tradeoffs of each technology allows us to save precious time in research and development and therefore to accelerate the time-to-market.

show abstract

“…Introduction. The phase change memory (PCM) is an attractive solution for embedded memories in automotive applications, due to its CMOS process compatibility, BEOL integration and good data retention at high temperature thanks to the optimized Ge-rich GST active material [1]. Both programmed states, namely the poly-crystalline set state and the amorphous reset state, show similar evolution, including a T-activated drift [2] according to the power law R/R(t0) = (t/t0)  , followed by a R drop due to crystallization and grain growth [3].…”

mentioning

confidence: 99%

“…Experiments. A wall-type PCM in a state-of-the-art bipolar-CMOS-DMOS (BCD) technology was studied in this work [1]. Fig.…”

mentioning

confidence: 99%