4:00 PM - EN13.06.02/EN14.14.02
Modeling the Temperature Dependent Seebeck Coefficient of Metastable Amorphous Ge2Sb2Te5 and Impacts of Thermoelectric Effects on the Operation of Phase Change Memory Devices
Helena Silva1,Jake Scoggin1,Noah Del Coro1,Md Tashfiq Bin Kashem1,Sadid Muneer1,Ali Gokirmak1
University of Connecticut1
Thermoelectric effects play a significant role in phase change memory (PCM) and Ovonic threshold switch (OTS) devices used as access devices in PCM cells . Typical PCM cells are two-terminal nanometer-scale resistive memory devices which can be reversibly switched between a low-resistance crystalline and a high-resistance amorphous state via nanosecond electrical pulses. Amorphization in PCM devices is achieved by self-heating the phase change material, typically a chalcogenide, close to its melting temperature, followed by a sudden quench. Crystallization is achieved by self-heating the phase change material above its glass-transition temperature. OTS devices typically use amorphous chalcogenides that do not crystallize during normal device operation.
The local current densities in PCM and OTS devices can reach 108 A / cm2, giving rise to local temperatures in excess of 900 K and thermal gradients as high as 50 K / nm; hence, Peltier effects at material interfaces and Thomson heating within the active area are substantial. Accurate modeling of thermoelectric effects requires knowledge of temperature dependent electrical resistivity and Seebeck coefficients of these materials. These parameters can be measured at low temperatures on as-deposited amorphous films . However, melt-quenched amorphous materials’ parameters tend to differ from as-deposited films, and PCM materials rapidly crystallize at higher temperatures. High-speed metastable electrical resistivity measurements can be performed on nanoscale devices using electrical pulses to uniformly amorphize devices up to approximately 200 nm in diameter , but larger devices tend to form current filaments and hence do not amorphize uniformly. On the other hand, measurement of the Seebeck coefficient (S), which is vital to understanding thermoelectric effects, is very difficult at small scales.
In this work, we model the Seebeck coefficient for metastable amorphous Ge2Sb2Te5 (aGST) based on high-speed experimental results  and an energy band diagram proposed by Muneer et. al.  from 300-850 K , , and we analyze thermoelectric effects in PCM cells using finite element phase change device simulations –. We calculate the electron and hole Seebeck contributions Se and Sh in metastable aGST with the band diagram in  and find that Sh is similar in both magnitude and slope to S measurements on as-deposited aGST thin films in 300-400 K range , , consistent with the unipolar conduction assumed when deriving the band gap in . We use Sh as the Seebeck coefficient in metastable aGST and simulate reset and set operations in a PCM double mushroom cell and find that the Seebeck differential between crystalline and amorphous GST results in significant heating/cooling at amorphous-crystalline junctions during both crystallization (set) and melting (reset).
 A. Faraclas et al., IEEE Electron Device Lett., 32, 12, pp. 1737–1739, Dec. (2011) DOI: 10.1109/LED.2011.2168374.
 L. Adnane et al., J. Appl. Phys., 122, 12, p. 125104, Sep. (2017) DOI: 10.1063/1.4996218.
 F. Dirisaglik et al., Nanoscale, 7, 40, pp. 16625–16630, Oct. (2015) DOI: 10.1039/C5NR05512A.
 S. Muneer et al., AIP Adv., 8, 6, (2018) DOI: 10.1063/1.5035085.
 L. Adnane et al., APS Meet., (2012).
 Z. Woods and A. Gokirmak, IEEE Trans. Electron Devices, 64, 11, pp. 4466–4471, Nov. (2017) DOI: 10.1109/TED.2017.2745506.
 Z. Woods et al., IEEE Trans. Electron Devices, 64, 11, pp. 4472–4478, Nov. (2017) DOI: 10.1109/TED.2017.2745500.
 J. Scoggin et al., Appl. Phys. Lett., 112, 19, p. 193502, (2018) DOI: 10.1063/1.5025331.
 J. Scoggin et al., Appl. Phys. Lett., 114, 4, pp. 1–6, Oct. (2019) DOI: 10.1063/1.5067397.
 S. A. Baily et al., Solid State Commun., 139, 4, pp. 161–164, (2006) DOI: 10.1016/j.ssc.2006.05.031.