arxiv: 2511.21213 · v2 · submitted 2025-11-26 · ❄️ cond-mat.mtrl-sci · cs.LG

Lattice-to-Total Thermal Conductivity Ratio: A Phonon-Glass Electron-Crystal Descriptor for Data-Driven Thermoelectric Design

Yifan Sun , Zhi Li , Tetsuya Imamura , Yuji Ohishi , Chris Wolverton , Ken Kurosaki This is my paper

Pith reviewed 2026-05-17 05:19 UTC · model grok-4.3

classification ❄️ cond-mat.mtrl-sci cs.LG

keywords thermoelectricsthermal conductivity ratiophonon-glass electron-crystalPGECmachine learning screeninghigh-ZT materialslattice thermal conductivitymaterial design

0 comments p. Extension

The pith

High-ZT thermoelectrics cluster near a lattice-to-total thermal conductivity ratio of 0.5.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper shows that high figure-of-merit thermoelectric materials appear not only in the low thermal conductivity regime but also cluster when the lattice portion is roughly half the total thermal conductivity. This ratio of approximately 0.5 supplies a concrete, measurable target that makes the phonon-glass electron-crystal design idea quantitative rather than qualitative. Machine learning models trained on 71,913 known entries then predict both total and lattice thermal conductivity, allowing joint screening of 104,567 compounds for candidates that are ultralow in conductivity and close to the target ratio. A doping example illustrates how the same models can suggest chemical changes that move a material toward the optimal ratio.

Core claim

Using a curated dataset of 71,913 entries, high-ZT materials cluster near κ_L/κ ≈ 0.5. This optimal ratio supplies a quantitative descriptor for the phonon-glass electron-crystal design concept. Two machine learning models, one for the lattice component and one for the electronic component of thermal conductivity, together yield both total κ and the ratio κ_L/κ. Applied to 104,567 inorganic compounds, the models identify 2,522 ultralow-κ candidates while scoring their proximity to the PGEC optimum. A follow-up doping case study shows how the framework can indicate chemical adjustments that shift pristine compounds closer to the κ_L/κ ≈ 0.5 target.

What carries the argument

The lattice-to-total thermal conductivity ratio (κ_L/κ) near 0.5, treated as a quantitative descriptor for the phonon-glass electron-crystal (PGEC) concept.

If this is right

High-ZT materials can be located by screening for both low total thermal conductivity and a lattice-to-total ratio near 0.5.
Joint machine-learning prediction of lattice and electronic thermal conductivity enables rapid evaluation of 100,000-scale compound libraries for both low κ and PGEC proximity.
Chemical doping can be directed to adjust the lattice and electronic contributions so that a given compound moves toward the ideal ratio of 0.5.
Discovery and performance optimization steps are combined in one framework that outputs both candidate lists and concrete tuning strategies.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

If the ratio proves causal, similar lattice-to-total descriptors might be tested in other systems where phonon scattering and electron transport must be decoupled.
Large-scale synthesis campaigns could prioritize compounds predicted to lie near 0.5 and then measure whether their ZT exceeds that of equally low-κ materials far from the ratio.
The same two-model approach could be retrained on experimental rather than computed thermal conductivity data to reduce the gap between prediction and measured performance.

Load-bearing premise

The clustering of high-ZT materials near κ_L/κ ≈ 0.5 arises from a causal design principle rather than from dataset biases or correlations with other material properties.

What would settle it

Measuring the ratio κ_L/κ for many newly synthesized high-ZT compounds and finding no statistical preference for values near 0.5 would undermine the claim that the ratio functions as a useful design descriptor.

Figures

Figures reproduced from arXiv: 2511.21213 by Chris Wolverton, Ken Kurosaki, Tetsuya Imamura, Yifan Sun, Yuji Ohishi, Zhi Li.

**Figure 1.** Figure 1: Relationships between ZT and thermal conductivity in the literature from Starrydata. (a) shows the relation between ZT, κL, and κe at approximately fixed total κ. (b) shows ZT versus κL/κ (lattice-to-total thermal conductivity ratio) for the full curated dataset. To illustrate more concretely how the tendency toward κL/κ → 0.5 reflects optimization toward the PGEC regime, we next examine the distribution o… view at source ↗

**Figure 2.** Figure 2: Relationships between ZT and the lattice-to-total thermal conductivity ratio (κL/κ), for two representative TE material families: (a) CoSb3 and (b) GeTe. Pristine compositions, defined as those whose reduced formula exactly matches the corresponding stoichiometric compound (CoSb3 or GeTe), are shown as filled crosses, while optimized variants are shown as semi-transparent circles. large κe [51, 52]. Howeve… view at source ↗

**Figure 3.** Figure 3: Fine-tuned model’s performance on the held-out test set for lattice thermal conductivity [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

**Figure 4.** Figure 4: Performance on the held-out test set for total thermal conductivity [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗

**Figure 5.** Figure 5: SHAP beeswarm plots for the (a) κL and (b) κe Random Forest models, computed on the entire training set. Features are ordered by mean |SHAP|. For the κL model, the unfilled d-electron count, mean NdUnfilled, has an overall positive effect on κL, meaning that compositions enriched in s-block and early d-block elements are predicted to have higher κL. This reflects the tendency of these elements to form stro… view at source ↗

**Figure 6.** Figure 6: (a) Workflow used to screen pristine compositions with ultralow total thermal conductivity at 300 K. (b) [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗

**Figure 7.** Figure 7: Predicted optimization paths for AgBiS2 (left), In4SnSe4 (middle), and TbCuTe2 (right) from the perspective of reducing κ while tuning κL/κ toward 0.5. Stars and open circles correspond to pristine and doped materials, respectively, and the dashed lines illustrate how selected dopants move κ and κL/κ. The green arrow highlights the PGEC design goal of achieving κL/κ ≈ 0.5 while lowering κ. For n-type AgBiS… view at source ↗

read the original abstract

Thermoelectrics (TEs) are promising candidates for energy harvesting with performance quantified by figure of merit, $ZT$. To accelerate the discovery of high-$ZT$ materials, efforts have focused on identifying compounds with low thermal conductivity $\kappa$. Using a curated dataset of 71,913 entries, we show that high-$ZT$ materials reside not only in the low-$\kappa$ regime but also cluster near a lattice-to-total thermal conductivity ratio ($\kappa_\mathrm{L}/\kappa$) of approximately 0.5. This optimal ratio provides a quantitative descriptor for the well-known phonon-glass electron-crystal (PGEC) design concept. Building on this insight, we construct a framework consisting of two machine learning models for the lattice and electronic components of thermal conductivity that jointly provide both $\kappa$ and $\kappa_\mathrm{L}/\kappa$ for screening and guiding the optimization of TE materials. By applying this framework to 104,567 inorganic compounds, we identify 2,522 ultralow-$\kappa$ candidates while simultaneously evaluating their proximity to the optimal PGEC regime. A follow-up case study on chemical doping demonstrates how the framework can qualitatively provide optimization strategies that shift pristine materials toward the ideal $\kappa_\mathrm{L}/\kappa$ $\approx$ 0.5 target. Ultimately, by integrating rapid screening with PGEC-guided optimization, our data-driven framework takes a critical step towards closing the gap between materials discovery and performance enhancement.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This work quantifies the PGEC concept with a κ_L/κ ≈ 0.5 target from dataset analysis and deploys ML models to screen compounds accordingly, though the target may reflect selection biases in the training data.

read the letter

The paper's main point is that in a large collection of thermoelectric data, the best performers sit near a lattice-to-total thermal conductivity ratio of about 0.5, and they use that to build a screening tool with machine learning. What stands out is the size of the effort. They pulled together over 71,000 data points and spotted this pattern in the high-ZT materials. From there they trained separate models for the lattice and electronic parts of thermal conductivity. That lets them predict both the total κ and the ratio for a much larger set of 104,000 compounds, pulling out a few thousand that look promising on both low conductivity and the PGEC-like ratio. The doping example at the end gives a concrete way to think about tweaking a material to move closer to that 0.5 mark. The work is solid on the practical side. Having two models instead of one for total κ is a nice touch because it directly gives the ratio they care about. Applying it at scale and showing how it could inform optimization is useful for anyone running high-throughput searches. Still, there are some things to watch. The clustering around 0.5 might not be a deep principle. It could just reflect that the dataset mostly includes materials that have already been optimized by doping or other means so that the electronic and lattice contributions end up similar. If that's the case, then using the same data to train the models and then screening new ones risks carrying that bias forward. The abstract does not spell out the validation numbers or baseline comparisons for the ML parts, so those details will matter a lot for how much weight to give the results. This kind of paper is for groups that combine data mining with thermoelectric materials research. Someone looking for new screening descriptors or ways to incorporate the PGEC idea into automated workflows would find it relevant. It has enough scale and a clear application angle that it should go to peer review rather than get desk rejected. I would send it out for review, but ask the referees to pay special attention to whether the 0.5 target is robust or just an artifact of the data collection.

Referee Report

2 major / 2 minor

Summary. The paper analyzes a curated dataset of 71,913 thermoelectric entries to show that high-ZT materials cluster near a lattice-to-total thermal conductivity ratio (κ_L/κ) of approximately 0.5, framing this as a quantitative descriptor for the phonon-glass electron-crystal (PGEC) concept. It develops two machine learning models to predict κ_L and κ_e (hence total κ and the ratio), screens 104,567 inorganic compounds to identify 2,522 ultralow-κ candidates near the target ratio, and presents a chemical-doping case study illustrating optimization toward κ_L/κ ≈ 0.5.

Significance. If the observed clustering proves robust and independent of dataset selection effects, the work supplies a concrete, data-driven extension of the PGEC heuristic that can be directly incorporated into high-throughput screening pipelines. The joint prediction of κ and the ratio, combined with the large-scale application and doping demonstration, offers a practical bridge between discovery and performance tuning. The manuscript's strength lies in its scale and the attempt to make PGEC actionable, though the overall significance is tempered by the need to rule out confounding variables in the central observational claim.

major comments (2)

[Dataset analysis section] Dataset analysis section (near the discussion of the 71,913-entry curation and Figure showing κ_L/κ distributions): the central claim that high-ZT materials cluster at κ_L/κ ≈ 0.5 as an independent PGEC optimum is load-bearing, yet the manuscript does not report controls for carrier concentration, doping level, or measurement temperature. Materials with both κ_L and total κ reported are often experimentally optimized TE compounds in which κ_e has been deliberately tuned via doping to approach κ_L; without subgroup analysis or regression controlling for these variables, the peak at 0.5 may reflect selection bias rather than a fundamental design principle.
[Machine learning framework and screening section] Machine learning framework and screening section (description of the two models and application to 104,567 compounds): the models for κ_L and κ_e are used to evaluate proximity to the 0.5 ratio for candidate selection, but no test-set MAE, R², or baseline comparisons (e.g., against composition-only regression or existing TE property predictors) are provided. This omission is critical because any systematic bias in the training data will propagate directly into the 2,522 identified candidates and their assigned ratios.

minor comments (2)

[Abstract] The abstract states the screened set as 104,567; ensure this exact figure and the 71,913 training-set size are consistently reported with commas in all tables and text for readability.
[Case study section] In the case-study section on chemical doping, clarify whether the predicted shifts in κ_L/κ are obtained by direct model inference on doped compositions or by some interpolation; the current description leaves the exact workflow ambiguous.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive and detailed comments, which highlight important aspects for strengthening the central claims and validation of our work. We address each major comment point by point below, indicating where revisions will be made to the manuscript.

read point-by-point responses

Referee: [Dataset analysis section] Dataset analysis section (near the discussion of the 71,913-entry curation and Figure showing κ_L/κ distributions): the central claim that high-ZT materials cluster at κ_L/κ ≈ 0.5 as an independent PGEC optimum is load-bearing, yet the manuscript does not report controls for carrier concentration, doping level, or measurement temperature. Materials with both κ_L and total κ reported are often experimentally optimized TE compounds in which κ_e has been deliberately tuned via doping to approach κ_L; without subgroup analysis or regression controlling for these variables, the peak at 0.5 may reflect selection bias rather than a fundamental design principle.

Authors: We agree that explicit controls for potential confounding variables are necessary to establish the robustness of the observed clustering. In the revised manuscript, we will add subgroup analyses stratified by reported measurement temperature ranges and, where metadata permits, by doping or carrier concentration levels. We will also include a multivariate regression analysis to assess the dependence of the κ_L/κ ratio on these factors while holding ZT fixed. This will allow us to evaluate whether the peak near 0.5 persists as an independent feature or is partly attributable to experimental optimization biases in the curated dataset. revision: yes
Referee: [Machine learning framework and screening section] Machine learning framework and screening section (description of the two models and application to 104,567 compounds): the models for κ_L and κ_e are used to evaluate proximity to the 0.5 ratio for candidate selection, but no test-set MAE, R², or baseline comparisons (e.g., against composition-only regression or existing TE property predictors) are provided. This omission is critical because any systematic bias in the training data will propagate directly into the 2,522 identified candidates and their assigned ratios.

Authors: We concur that quantitative model validation metrics are essential for assessing the reliability of the screening results and for identifying any propagated biases. In the revised manuscript, we will report test-set MAE and R² values for both the κ_L and κ_e models, along with details of the train-test split and cross-validation procedure. We will additionally include baseline comparisons against composition-only regression models and selected existing thermoelectric property predictors from the literature to quantify performance gains and to discuss implications for the 2,522 candidates. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical observation used as descriptor without reduction to inputs by construction

full rationale

The paper curates a dataset of 71,913 entries and reports an empirical clustering of high-ZT materials near κ_L/κ ≈ 0.5 as an observed pattern that quantifies the PGEC concept. ML models are then trained on this data to predict κ_L and κ_e for screening 104,567 compounds and evaluating proximity to the observed ratio. No equation, derivation, or self-citation reduces the claimed descriptor or screening result to a fitted parameter or prior result by construction. The 0.5 value is presented as a data-driven finding rather than a self-defined target or renamed known result, and the framework remains self-contained as an empirical analysis with independent predictive components.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The framework rests on the assumption that the observed ratio is a useful design handle and that ML models trained on existing thermoelectric data can be applied to new inorganic compounds without large domain-shift errors.

free parameters (1)

optimal κ_L/κ ratio
The value 0.5 is chosen because high-ZT entries cluster near it in the 71k dataset; it functions as a fitted target rather than a first-principles constant.

axioms (1)

domain assumption The phonon-glass electron-crystal (PGEC) concept is a valid guiding principle for thermoelectric optimization
Invoked to interpret the observed ratio as a design target.

pith-pipeline@v0.9.0 · 5592 in / 1356 out tokens · 41379 ms · 2026-05-17T05:19:30.877914+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel echoes

?

echoes
ECHOES: this paper passage has the same mathematical shape or conceptual pattern as the Recognition theorem, but is not a direct formal dependency.

high-ZT materials ... cluster near a lattice-to-total thermal conductivity ratio (κ_L/κ) of approximately 0.5. This optimal ratio provides a quantitative descriptor for the well-known phonon-glass electron-crystal (PGEC) design concept.

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

64 extracted references · 64 canonical work pages

[1]

Estimating the global waste heat potential.Renewable and Sustainable Energy Reviews, 57:1568–1579, 2016

Clemens Forman, Ibrahim Kolawole Muritala, Robert Pardemann, and Bernd Meyer. Estimating the global waste heat potential.Renewable and Sustainable Energy Reviews, 57:1568–1579, 2016

work page 2016
[2]

Techno-economic analysis of waste-heat conversion.Joule, 5(12):3080–3096, 2021

Charles Geffroy, Drew Lilley, Pedro Sanchez Parez, and Ravi Prasher. Techno-economic analysis of waste-heat conversion.Joule, 5(12):3080–3096, 2021

work page 2021
[3]

Cooling, heating, generating power, and recovering waste heat with thermoelectric systems.Science, 321(5895):1457–1461, 2008

Lon E Bell. Cooling, heating, generating power, and recovering waste heat with thermoelectric systems.Science, 321(5895):1457–1461, 2008

work page 2008
[4]

Complex thermoelectric materials.Nature Materials, 7(2):105–114, 2008

G Jeffrey Snyder and Eric S Toberer. Complex thermoelectric materials.Nature Materials, 7(2):105–114, 2008

work page 2008
[5]

A public database of thermoelectric materials and system-identified material representation for data-driven discovery.npj Computational Materials, 8(1):214, 2022

Gyoung S Na and Hyunju Chang. A public database of thermoelectric materials and system-identified material representation for data-driven discovery.npj Computational Materials, 8(1):214, 2022

work page 2022
[6]

Large data set-driven machine learning models for accurate prediction of the thermoelectric figure of merit.ACS Applied Materials & Interfaces, 14(50): 55517–55527, 2022

Yi Li, Jingzi Zhang, Ke Zhang, Mengkun Zhao, Kailong Hu, and Xi Lin. Large data set-driven machine learning models for accurate prediction of the thermoelectric figure of merit.ACS Applied Materials & Interfaces, 14(50): 55517–55527, 2022

work page 2022
[7]

Machine learning for predicting ZT values of high-performance thermoelectric materials in mid-temperature range.APL Materials, 11(8), 2023

Nuttawat Parse and Supree Pinitsoontorn. Machine learning for predicting ZT values of high-performance thermoelectric materials in mid-temperature range.APL Materials, 11(8), 2023. 12 A PREPRINT

work page 2023
[8]

Thermoelectric material performance (zT) predictions with machine learning.ACS Applied Materials & Interfaces, 17(1):1662–1673, 2024

Nikhil K Barua, Sangjoon Lee, Anton O Oliynyk, and Holger Kleinke. Thermoelectric material performance (zT) predictions with machine learning.ACS Applied Materials & Interfaces, 17(1):1662–1673, 2024

work page 2024
[9]

Prediction of thermoelectric-figure-of-merit based on autoencoder and light gradient boosting machine.Journal of Applied Physics, 135(7), 2024

Yingying Xu, Xinyi Liu, and Jifen Wang. Prediction of thermoelectric-figure-of-merit based on autoencoder and light gradient boosting machine.Journal of Applied Physics, 135(7), 2024

work page 2024
[10]

Prediction of seebeck coefficient for compounds without restriction to fixed stoichiometry: A machine learning approach.Journal of Computational Chemistry, 39(4):191–202, 2018

Al’ona Furmanchuk, James E Saal, Jeff W Doak, Gregory B Olson, Alok Choudhary, and Ankit Agrawal. Prediction of seebeck coefficient for compounds without restriction to fixed stoichiometry: A machine learning approach.Journal of Computational Chemistry, 39(4):191–202, 2018

work page 2018
[11]

Machine learning for accelerated prediction of the seebeck coefficient at arbitrary carrier concentration.Materials Today Physics, 25:100706, 2022

HM Yuan, SH Han, R Hu, WY Jiao, MK Li, HJ Liu, and Y Fang. Machine learning for accelerated prediction of the seebeck coefficient at arbitrary carrier concentration.Materials Today Physics, 25:100706, 2022

work page 2022
[12]

Machine learning predictions of thermopower for thermoelectric material screening.ACS Applied Energy Materials, 2025

Nikhil K Barua and Holger Kleinke. Machine learning predictions of thermopower for thermoelectric material screening.ACS Applied Energy Materials, 2025

work page 2025
[13]

Machine learning approaches to identify and design low thermal conductivity oxides for thermoelectric applications.Data-Centric Engineering, 1: e8, 2020

Abhishek Tewari, Siddharth Dixit, Niteesh Sahni, and Stéphane PA Bordas. Machine learning approaches to identify and design low thermal conductivity oxides for thermoelectric applications.Data-Centric Engineering, 1: e8, 2020

work page 2020
[14]

Identification of crystalline materials with ultra-low thermal conductivity based on machine learning study.The Journal of Physical Chemistry C, 124(16):8488–8495, 2020

Xinming Wang, Shuming Zeng, Zhuchi Wang, and Jun Ni. Identification of crystalline materials with ultra-low thermal conductivity based on machine learning study.The Journal of Physical Chemistry C, 124(16):8488–8495, 2020

work page 2020
[15]

Interpretable machine learning model on thermal conductivity using publicly available datasets and our internal lab dataset.Chemistry of Materials, 36(14):7089–7100, 2024

Nikhil K Barua, Evan Hall, Yifei Cheng, Anton O Oliynyk, and Holger Kleinke. Interpretable machine learning model on thermal conductivity using publicly available datasets and our internal lab dataset.Chemistry of Materials, 36(14):7089–7100, 2024

work page 2024
[16]

Machine learning for accelerated prediction of lattice thermal conductivity at arbitrary temperature.Digital Discovery, 4(1):204–210, 2025

Zihe Li, Mengke Li, Yufeng Luo, Haibin Cao, Huijun Liu, and Ying Fang. Machine learning for accelerated prediction of lattice thermal conductivity at arbitrary temperature.Digital Discovery, 4(1):204–210, 2025

work page 2025
[17]

Defect engineering for realizing high thermoelectric performance in n-type Mg3Sb2-based materials.ACS Energy Letters, 2(10):2245–2250, 2017

Jun Mao, Yixuan Wu, Shaowei Song, Qing Zhu, Jing Shuai, Zihang Liu, Yanzhong Pei, and Zhifeng Ren. Defect engineering for realizing high thermoelectric performance in n-type Mg3Sb2-based materials.ACS Energy Letters, 2(10):2245–2250, 2017

work page 2017
[18]

Defect-engineering-stabilized AgSbTe2 with high thermoelectric performance

Yu Zhang, Zhi Li, Saurabh Singh, Amin Nozariasbmarz, Wenjie Li, Aziz Genç, Yi Xia, Luyao Zheng, Seng Huat Lee, Sumanta Kumar Karan, et al. Defect-engineering-stabilized AgSbTe2 with high thermoelectric performance. Advanced Materials, 35(11):2208994, 2023

work page 2023
[19]

Integrating band structure engineering with all-scale hierarchical structuring for high thermoelectric performance in PbTe system.Advanced Energy Materials, 7(3):1601450, 2017

Yanling Pei, Gangjian Tan, Dan Feng, Lei Zheng, Qing Tan, Xiaobing Xie, Shengkai Gong, Yue Chen, Jing-Feng Li, Jiaqing He, et al. Integrating band structure engineering with all-scale hierarchical structuring for high thermoelectric performance in PbTe system.Advanced Energy Materials, 7(3):1601450, 2017

work page 2017
[20]

Hongyao Xie, Yukun Liu, Yinying Zhang, Shiqiang Hao, Zhi Li, Matthew Cheng, Songting Cai, G Jeffrey Snyder, Christopher Wolverton, Ctirad Uher, et al. High Thermoelectric Performance in Chalcopyrite Cu1-xAgx GaTe2–ZnTe: Nontrivial Band Structure and Dynamic Doping Effect.Journal of the American Chemical Society, 144(20):9113–9125, 2022

work page 2022
[21]

High wide-temperature-range thermoelectric performance in GeTe through hetero- nanostructuring.Acta Materialia, 276:120132, 2024

Qingtang Zhang, Pan Ying, Aftab Farrukh, Yaru Gong, Jizi Liu, Xinqi Huang, Di Li, Meiyu Wang, Guang Chen, and Guodong Tang. High wide-temperature-range thermoelectric performance in GeTe through hetero- nanostructuring.Acta Materialia, 276:120132, 2024

work page 2024
[22]

Yun Zheng, Qiang Zhang, Xianli Su, Hongyao Xie, Shengcheng Shu, Tianle Chen, Gangjian Tan, Yonggao Yan, Xinfeng Tang, Ctirad Uher, et al. Mechanically robust BiSbTe alloys with superior thermoelectric performance: a case study of stable hierarchical nanostructured thermoelectric materials.Advanced Energy Materials, 5(5): 1401391, 2015

work page 2015
[23]

CRC handbook of thermoelectrics, 1995

Glen A Slack, DM Rowe, et al. CRC handbook of thermoelectrics, 1995

work page 1995
[24]

Skutterudites: A phonon-glass-electron crystal approach to advanced thermoelectric energy conversion applications.Annual Review of Materials Science, 29(1):89–116, 1999

GS Nolas, DT Morelli, and Terry M Tritt. Skutterudites: A phonon-glass-electron crystal approach to advanced thermoelectric energy conversion applications.Annual Review of Materials Science, 29(1):89–116, 1999

work page 1999
[25]

Phonon-glass electron-crystal thermoelectric clathrates: Experiments and theory.Reviews of Modern Physics, 86(2):669–716, 2014

Toshiro Takabatake, Koichiro Suekuni, Tsuneyoshi Nakayama, and Eiji Kaneshita. Phonon-glass electron-crystal thermoelectric clathrates: Experiments and theory.Reviews of Modern Physics, 86(2):669–716, 2014

work page 2014
[26]

Disordered zinc in Zn4Sb3 with phonon-glass and electron-crystal thermoelectric properties.Nature Materials, 3(7):458–463, 2004

G Jeffrey Snyder, Mogens Christensen, Eiji Nishibori, Thierry Caillat, and Bo Brummerstedt Iversen. Disordered zinc in Zn4Sb3 with phonon-glass and electron-crystal thermoelectric properties.Nature Materials, 3(7):458–463, 2004

work page 2004
[27]

Starrydata: from published plots to shared materials data

Yukari Katsura, Masaya Kumagai, Tomoya Mato, Yu Takada, Yuki Ando, Erina Fujita, Fumikazu Hosono, Eiji Koyama, Farhan Mudasar, Ton Nu Thanh Phuong, et al. Starrydata: from published plots to shared materials data. Science and Technology of Advanced Materials: Methods, 5(1):2506976, 2025. 13 A PREPRINT

work page 2025
[28]

A general-purpose machine learning framework for predicting properties of inorganic materials.npj Computational Materials, 2(1):1–7, 2016

Logan Ward, Ankit Agrawal, Alok Choudhary, and Christopher Wolverton. A general-purpose machine learning framework for predicting properties of inorganic materials.npj Computational Materials, 2(1):1–7, 2016

work page 2016
[29]

Commentary: The Materials Project: A materials genome approach to accelerating materials innovation.APL Materials, 1(1), 2013

Anubhav Jain, Shyue Ping Ong, Geoffroy Hautier, Wei Chen, William Davidson Richards, Stephen Dacek, Shreyas Cholia, Dan Gunter, David Skinner, Gerbrand Ceder, et al. Commentary: The Materials Project: A materials genome approach to accelerating materials innovation.APL Materials, 1(1), 2013

work page 2013
[30]

Characterization of Lorenz number with Seebeck coefficient measurement.APL Materials, 3(4), 2015

Hyun-Sik Kim, Zachary M Gibbs, Yinglu Tang, Heng Wang, and G Jeffrey Snyder. Characterization of Lorenz number with Seebeck coefficient measurement.APL Materials, 3(4), 2015

work page 2015
[31]

Copper ion liquid-like thermoelectrics.Nature materials, 11(5):422–425, 2012

Huili Liu, Xun Shi, Fangfang Xu, Linlin Zhang, Wenqing Zhang, Lidong Chen, Qiang Li, Ctirad Uher, Tristan Day, and G Jeffrey Snyder. Copper ion liquid-like thermoelectrics.Nature materials, 11(5):422–425, 2012

work page 2012
[32]

Lisha Xue, Weixia Shen, Zhuangfei Zhang, Manjie Shen, Wenting Ji, Chao Fang, Yuewen Zhang, and Xiaopeng Jia. Off-stoichiometry effects on the thermoelectric properties of Cu2+δSe (-0.1≤δ≤ 0.05) compounds synthesized by a high-pressure and high-temperature method.CrystEngComm, 22(4):695–700, 2020

work page 2020
[33]

Enhanced performances of melt spun Bi 2(Te,Se)3 for n-type thermoelectric legs.Intermetallics, 19(7):1024–1031, 2011

Shanyu Wang, Wenjie Xie, Han Li, and Xinfeng Tang. Enhanced performances of melt spun Bi 2(Te,Se)3 for n-type thermoelectric legs.Intermetallics, 19(7):1024–1031, 2011

work page 2011
[34]

Bing Sun, Xiaopeng Jia, Dexuan Huo, Hairui Sun, Yuewen Zhang, Binwu Liu, Haiqiang Liu, Lingjiao Kong, Baomin Liu, and Hongan Ma. Effect of high-temperature and high-pressure processing on the structure and thermoelectric properties of clathrate Ba8Ga16Ge30.The Journal of Physical Chemistry C, 120(18):10104–10110, 2016

work page 2016
[35]

High-temperature thermoelectric properties of thallium-filled skutterudites.Applied Physics Letters, 96(20), 2010

Adul Harnwunggmoung, Ken Kurosaki, Hiroaki Muta, and Shinsuke Yamanaka. High-temperature thermoelectric properties of thallium-filled skutterudites.Applied Physics Letters, 96(20), 2010

work page 2010
[36]

Enhanced thermoelectric properties of Bi 0.5Sb1.5Te3 composites with in-situ formed senarmontite Sb2O3 nanophase.Journal of Alloys and Compounds, 777:703–711, 2019

Eun Bin Kim, Peyala Dharmaiah, Kap-Ho Lee, Chul-Hee Lee, Jong-Hyeon Lee, Jae-Kyo Yang, Dae-Hwan Jang, Dong-Soo Kim, and Soon-Jik Hong. Enhanced thermoelectric properties of Bi 0.5Sb1.5Te3 composites with in-situ formed senarmontite Sb2O3 nanophase.Journal of Alloys and Compounds, 777:703–711, 2019

work page 2019
[37]

Beneficial influence of iodine substitution on the thermoelectric properties of Mo3Sb7.Journal of Applied Physics, 127(10), 2020

Sora-at Tanusilp, Suphagrid Wongprakarn, Pinit Kidkhunthod, Yuji Ohishi, Hiroaki Muta, and Ken Kurosaki. Beneficial influence of iodine substitution on the thermoelectric properties of Mo3Sb7.Journal of Applied Physics, 127(10), 2020

work page 2020
[38]

Substantial thermoelectric enhancement achieved by manipulating the band structure and dislocations in Ag and La co-doped SnTe.Journal of Advanced Ceramics, 10(4):860–870, 2021

Wenjing Xu, Zhongwei Zhang, Chengyan Liu, Jie Gao, Zhenyuan Ye, Chunguang Chen, Ying Peng, Xiaobo Bai, and Lei Miao. Substantial thermoelectric enhancement achieved by manipulating the band structure and dislocations in Ag and La co-doped SnTe.Journal of Advanced Ceramics, 10(4):860–870, 2021

work page 2021
[39]

Unraveling the thermoelectric performance of Bismuth Antimony/graphene nanocomposite synthesized by spark plasma extrusion.Journal of Alloys and Compounds, 887:161399, 2021

Mohamed S El-Asfoury, Shaban M Abdou, and Ahmed Nassef. Unraveling the thermoelectric performance of Bismuth Antimony/graphene nanocomposite synthesized by spark plasma extrusion.Journal of Alloys and Compounds, 887:161399, 2021

work page 2021
[40]

Xin Liang and Chuang Chen. Ductile inorganic amorphous/crystalline composite Ag 4TeS with phonon-glass electron-crystal transport behavior and excellent stability of high thermoelectric performance on plastic deforma- tion.Acta Materialia, 218:117231, 2021

work page 2021
[41]

Medium-temperature thermoelectric GeTe: vacancy suppression and band structure engineering leading to high performance.Energy & Environmental Science, 12(4):1396–1403, 2019

Jinfeng Dong, Fu-Hua Sun, Huaichao Tang, Jun Pei, Hua-Lu Zhuang, Hai-Hua Hu, Bo-Ping Zhang, Yu Pan, and Jing-Feng Li. Medium-temperature thermoelectric GeTe: vacancy suppression and band structure engineering leading to high performance.Energy & Environmental Science, 12(4):1396–1403, 2019

work page 2019
[42]

Tin Telluride-Based Nanocomposites of the Type AgSn mBiTe2+m (BTST-m) as Effective Lead-Free Thermoelectric Materials.Chemistry of Materials, 27(21):7296–7305, 2015

Oliver Falkenbach, Andreas Schmitz, Torben Dankwort, Guenter Koch, Lorenz Kienle, Eckhard Mueller, and Sabine Schlecht. Tin Telluride-Based Nanocomposites of the Type AgSn mBiTe2+m (BTST-m) as Effective Lead-Free Thermoelectric Materials.Chemistry of Materials, 27(21):7296–7305, 2015

work page 2015
[43]

Ultrafast synthesis of Te-doped CoSb3 with excellent thermoelectric properties.ACS Applied Energy Materials, 2(6):4477–4485, 2019

Ying Lei, Wensheng Gao, Rui Zheng, Yu Li, Wen Chen, Libo Zhang, Rundong Wan, Hongwei Zhou, Zhiyuan Liu, and Paul K Chu. Ultrafast synthesis of Te-doped CoSb3 with excellent thermoelectric properties.ACS Applied Energy Materials, 2(6):4477–4485, 2019

work page 2019
[44]

Xun Shi, Jiong Yang, James R Salvador, Miaofang Chi, Jung Y Cho, Hsin Wang, Shengqiang Bai, Jihui Yang, Wenqing Zhang, and Lidong Chen. Multiple-filled skutterudites: high thermoelectric figure of merit through separately optimizing electrical and thermal transports.Journal of the American Chemical Society, 133(20): 7837–7846, 2011

work page 2011
[45]

Properties of single crystalline semiconducting CoSb3.Journal of Applied Physics, 80(8):4442–4449, 1996

T Caillat, A Borshchevsky, and J-P Fleurial. Properties of single crystalline semiconducting CoSb3.Journal of Applied Physics, 80(8):4442–4449, 1996

work page 1996
[46]

Thermoelectric properties of Yb-filled CoSb3 skutterudites.Journal of the Korean Physical Society, 65(4):491–495, 2014

Kwan-Ho Park, Won-Seon Seo, Dong-Kil Shin, and Il-Ho Kim. Thermoelectric properties of Yb-filled CoSb3 skutterudites.Journal of the Korean Physical Society, 65(4):491–495, 2014. 14 A PREPRINT

work page 2014
[47]

Y Dou, J Li, Y Xie, X Wu, L Hu, F Liu, W Ao, Y Liu, and C Zhang. Lone-pair engineering: Achieving ultralow lattice thermal conductivity and enhanced thermoelectric performance in Al-doped GeTe-based alloys.Materials Today Physics, 20:100497, 2021

work page 2021
[48]

Zihang Liu, Jifeng Sun, Jun Mao, Hangtian Zhu, Wuyang Ren, Jingchao Zhou, Zhiming Wang, David J Singh, Jiehe Sui, Ching-Wu Chu, et al. Phase-transition temperature suppression to achieve cubic GeTe and high thermoelectric performance by Bi and Mn codoping.Proceedings of the National Academy of Sciences, 115(21): 5332–5337, 2018

work page 2018
[49]

Simulta- neous optimization of carrier concentration and alloy scattering for ultrahigh performance GeTe thermoelectrics

Juan Li, Zhiwei Chen, Xinyue Zhang, Hulei Yu, Zihua Wu, Huaqing Xie, Yue Chen, and Yanzhong Pei. Simulta- neous optimization of carrier concentration and alloy scattering for ultrahigh performance GeTe thermoelectrics. Advanced Science, 4(12):1700341, 2017

work page 2017
[50]

Resonant level-induced high thermoelectric response in indium-doped GeTe.NPG Asia Materials, 9(1):e343–e343, 2017

Lihua Wu, Xin Li, Shanyu Wang, Tiansong Zhang, Jiong Yang, Wenqing Zhang, Lidong Chen, and Jihui Yang. Resonant level-induced high thermoelectric response in indium-doped GeTe.NPG Asia Materials, 9(1):e343–e343, 2017

work page 2017
[51]

Electronic and thermal transport in GeTe: A versatile base for thermoelectric materials.Journal of Applied Physics, 114(8), 2013

EM Levin, MF Besser, and Riley Hanus. Electronic and thermal transport in GeTe: A versatile base for thermoelectric materials.Journal of Applied Physics, 114(8), 2013

work page 2013
[52]

Realizing zT of 2.3 in Ge1−x−ySbxInyTe via reducing the phase-transition temperature and introducing resonant energy doping.Advanced materials, 30(11):1705942, 2018

Min Hong, Zhi-Gang Chen, Lei Yang, Yi-Chao Zou, Matthew S Dargusch, Hao Wang, and Jin Zou. Realizing zT of 2.3 in Ge1−x−ySbxInyTe via reducing the phase-transition temperature and introducing resonant energy doping.Advanced materials, 30(11):1705942, 2018

work page 2018
[53]

Realizing the high thermoelectric performance of GeTe by Sb-doping and Se-alloying.Chemistry of Materials, 29(2):605–611, 2017

Juan Li, Xinyue Zhang, Siqi Lin, Zhiwei Chen, and Yanzhong Pei. Realizing the high thermoelectric performance of GeTe by Sb-doping and Se-alloying.Chemistry of Materials, 29(2):605–611, 2017

work page 2017
[54]

Evaluation of the structure and transport properties of nanostructured antimony telluride (Sb2Te3).Journal of Electronic Materials, 43(6):1927–1932, 2014

Mohsin Saleemi, A Ruditskiy, MS Toprak, Marian Stingaciu, Mats Johnsson, I Kretzschmar, A Jacquot, M Jägle, and Mamoun Muhammed. Evaluation of the structure and transport properties of nanostructured antimony telluride (Sb2Te3).Journal of Electronic Materials, 43(6):1927–1932, 2014

work page 1927
[55]

Enhanced figure of merit in antimony telluride thermoelectric materials by In–Ag co-alloying for mid-temperature power generation.Acta Materialia, 85:270–278, 2015

LP Hu, TJ Zhu, XQ Yue, XH Liu, YG Wang, ZJ Xu, and XB Zhao. Enhanced figure of merit in antimony telluride thermoelectric materials by In–Ag co-alloying for mid-temperature power generation.Acta Materialia, 85:270–278, 2015

work page 2015
[56]

Heng Quan Yang, Lei Miao, Cheng Yan Liu, Chao Li, Sawao Honda, Yuji Iwamoto, Rong Huang, and Sakae Tanemura. A facile surfactant-assisted reflux method for the synthesis of single-crystalline Sb2Te3 nanostructures with enhanced thermoelectric performance.ACS Applied Materials & Interfaces, 7(26):14263–14271, 2015

work page 2015
[57]

Guo-Hui Dong, Ying-Jie Zhu, and Li-Dong Chen. Microwave-assisted rapid synthesis of Sb2Te3 nanosheets and thermoelectric properties of bulk samples prepared by spark plasma sintering.Journal of Materials Chemistry, 20 (10):1976–1981, 2010

work page 1976
[58]

Structured information extraction from scientific text with large language models

John Dagdelen, Alexander Dunn, Sanghoon Lee, Nicholas Walker, Andrew S Rosen, Gerbrand Ceder, Kristin A Persson, and Anubhav Jain. Structured information extraction from scientific text with large language models. Nature Communications, 15(1):1418, 2024

work page 2024
[59]

Large language model-driven database for thermoelectric materials

Suman Itani, Yibo Zhang, and Jiadong Zang. Large language model-driven database for thermoelectric materials. Computational Materials Science, 253:113855, 2025

work page 2025
[60]

Hongxu An, Dongyang Wang, and Wenke He. Thermoelectric properties of AgBiS 2: Unveiling the origin of ultra-low lattice thermal conductivity and optimization strategies for electrical performance.Applied Physics Letters, 127(14), 2025

work page 2025
[61]

Synthesis, structure, Te alloying, and physical properties of CuSbS2.Inorganic Chemistry, 56(22):14040–14044, 2017

Dean Hobbis, Kaya Wei, Hsin Wang, Joshua Martin, and George S Nolas. Synthesis, structure, Te alloying, and physical properties of CuSbS2.Inorganic Chemistry, 56(22):14040–14044, 2017

work page 2017
[62]

A promising thermoelectrics In 4SnSe4 with a wide bandgap and cubic structure composited by layered SnSe and In 4Se3

Haonan Shi, Changrong Guo, Bingchao Qin, Guangtao Wang, Dongyang Wang, and Li-Dong Zhao. A promising thermoelectrics In 4SnSe4 with a wide bandgap and cubic structure composited by layered SnSe and In 4Se3. Journal of Materiomics, 8(5):982–991, 2022

work page 2022
[63]

Thermoelectric Properties and Electronic Structure of the Zintl-Phase Sr3AlSb3.ChemSusChem, 6(12):2316–2321, 2013

Alex Zevalkink, Gregory Pomrehn, Yoshiki Takagiwa, Jessica Swallow, and G Jeffrey Snyder. Thermoelectric Properties and Electronic Structure of the Zintl-Phase Sr3AlSb3.ChemSusChem, 6(12):2316–2321, 2013

work page 2013
[64]

Syntheses, structures, and thermoelectric properties of ternary tellurides: RECuTe 2 (RE=Tb–Er).Inorganic Chemistry Frontiers, 4(8):1273–1280, 2017

Hua Lin, Hong Chen, Ni Ma, Yu-Jun Zheng, Jin-Ni Shen, Ju-Song Yu, Xin-Tao Wu, and Li-Ming Wu. Syntheses, structures, and thermoelectric properties of ternary tellurides: RECuTe 2 (RE=Tb–Er).Inorganic Chemistry Frontiers, 4(8):1273–1280, 2017. 15

work page 2017