Low‐Temperature Transport in Crystalline Ge₁Sb₂Te₄

Abstract

Disorder and its reduction upon annealing play a crucial role in understanding the electrical transport in the crystalline phase‐change material Ge1Sb2Te4. Previous studies focus either on the impact of disorder at moderate temperatures or on the low‐temperature properties of crystalline films with a low degree of disorder. The present investigation describes and discusses the impact of pronounced disorder on charge transport at low temperatures. The present data reveal the existence of a metal‐to‐insulator transition (MIT), where upon increasing order the zero‐temperature limit of conductivity changes from zero (insulator) to nonzero values (metal). The position of the MIT is determined with respect to the control parameter, i.e., the disorder, which is modified through the annealing conditions. Disorder is shown to localize carriers for an exceptionally large density of states. In the most disordered films, variable range hopping is observed, enabling the determination of the localization length. At the lowest temperatures studied, deviations from Mott variable range hopping are observed, which can be explained by a transition to Efros–Shklovskii hopping due to the presence of a soft Coulomb gap.

The metal-insulator transition in disordered solids: How theoretical prejudices influence its characterization A critical review of analyses of experimental data

“…Volker et al 47 not only extended the explored T range by one order of magnitude down to 0.35 K but also changed their interpretation approach: in locating the MIT, they attached particular importance to the s D a C b T 1/2 approximation, now trusting in it as "the most widely accepted extrapolation method". 47 Indeed, Figure 2 of Volker et al 47 shows that, again, this ansatz works rather well.…”

“…One and a half years after a first preliminary version of this review had been made available at arXiv.org, see M€ obius, 97 three of the authors of Siegrist et al 31 submitted a subsequent study of the MIT in GeSb 2 Te 4 , Volker et al 47 Apparently, in Siegrist et al 31 and Volker et al, 47 (almost) identical sample preparation procedures were used, compare the preparation details given in the Supplementary Information of Siegrist et al 31 and in Section 2 of Volker et al 47 with each other. Volker et al 47 not only extended the explored T range by one order of magnitude down to 0.35 K but also changed their interpretation approach: in locating the MIT, they attached particular importance to the s D a C b T 1/2 approximation, now trusting in it as "the most widely accepted extrapolation method".…”

“…The situation gets even more complicated when data from a subsequent GeSb 2 Te 4 study, Volker et al, 47 by three of the authors of Siegrist et al 31 are additionally taken into account. Therein, the T range was extended by one order of magnitude, and continuity of the MIT was concluded from augmented power law approximations of the measured s(T ).…”

“…In doing so, we demonstrate inconsistencies in the data sets from Siegrist et al 31 which render it impossible to reach definite conclusions about the nature of conduction for three of the samples. Moreover, we explain why a part of the sample classifications of Volker et al 47 seems to be incorrect, so that the characterization of the MIT therein is called into question. In the second part of this section, Subsection 3.2, several arguments against the hypothetical violation of the Mott criterion by GeSb 2 Te 4 are presented; this deviation is found to be not real but to result from an invalid assumption on the participating states.…”

“…(Readers in a hurry may focus on Subsections 2.2, 2.3, 2.4, and, in particular, 2.6.) Section 3 is devoted to GeSb 2 Te 4 : In its first part, Subsection 3.1, s(T) data from Siegrist et al 31 and Volker et al 47 are reanalyzed by means of alternative approaches. In doing so, we demonstrate inconsistencies in the data sets from Siegrist et al 31 which render it impossible to reach definite conclusions about the nature of conduction for three of the samples.…”