Mechanical Performance Database for Low-Temperature Alloys

Haoxuan Tang, Zhiyuan Chen, Xin Yao, Zhiping Xu · 2025

Read Mechanical Performance Database for Low-Temperature Alloys by Haoxuan Tang, Zhiyuan Chen, Xin Yao, Zhiping Xu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-13734.txt.

Introduction

Haoxuan Tang, Zhiyuan Chen, Xin Yao, and Zhiping Xu wrote this materials paper, published in two thousand twenty-five, to present a database of metals that work in deep cold. Liquid hydrogen tanks, superconductors, and quantum gear all need alloys that stay strong and tough when very cold.

The team mined published papers for test results, using machine reading to pull numbers, then checked and fixed them by hand. The records cover steels, aluminum and titanium alloys, and newer medium- and high-entropy alloys. Each entry lists yield and tensile strength, stretch before breaking, and impact energy, along with the alloy's recipe and test conditions. A summary of the data shows where research clusters and where it thins out. The database is posted in an open store for any builder or researcher to use.

Catalog record

CREATORHaoxuan Tang, Zhiyuan Chen, Xin Yao, Zhiping Xu
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE26 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-13734

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Mechanical Performance Database for Low-Temperature Alloys Haoxuan Tang, Zhiyuan Chen, Xin Yao, Zhiping Xu Abstract Low-temperature alloys are important for a wide spectrum of modern technologies ranging from liquid hydrogen, and superconductivity to quantum technology. These applications push the limit of material performance into extreme coldness, often demanding a combination of strength and toughness to address various challenges. Steel is one of the most widely used materials in cryogenic applications. With the deployment in aerospace liquid hydrogen storage and the transportation industry, aluminum and titanium alloys are also gaining increasing attention. Emerging medium-entropy alloys (MEAs) and high-entropy alloys (HEAs) demonstrate excellent low-temperature mechanical performance with a much-expanded space of material design. A database of low-temperature metallic alloys is collected from the literature and hosted in an open repository. The workflow of data collection includes automated extraction based on machine learning and natural language processing, supplemented by manual inspection and correction, to enhance data extraction efficiency and database quality. The product datasets cover key performance parameters including yield strength, tensile strength, elongation at fracture, Charpy impact energy, as well as detailed information on materials such as their types, chemical compositions, processing and testing conditions. Data statistics are analyzed to elucidate the research and development patterns and clarify the challenges in both scientific exploration and engineering deployment. keywords Metallic alloys; low-temperature applications; mechanical properties; research data; open science Background & Summary The term ‘cryogenics’ is defined for applications at T<120 K timmerhaus2007cryogenic . A wide spectrum of technologies is deployed in low-temperature (low-T) environments, such as quantum technologies including quantum computing and sensing, superconducting-enabled applications such as magnetic resonance imaging (MRI) ( 4.2 K), aerospace or transportation technologies powered by liquid hydrogen (LH2) fuels ( 20 K), and liquefied natural gas storage container ( 110 K) (Fig. 1 a). In these low and discrete temperature ranges, metallic alloys may behave differently from those at room temperature (RT) where material behaviors are better understood. In practice, achieving precise control of low T poses certain challenges for refrigeration techniques. In low-T experiments, materials are typically either immersed directly in the refrigerants or cooled by refrigeration systems. The latter allows for more accurate temperature control and can achieve even lower temperatures (on the order of milli-Kelvin) zu2022development (Fig. 1 d). OExtreme cold presents significant challenges for maintaining material performance, making material screening and selection critical in the design of low-temperature technologies. A standardized database with information on material properties and refrigeration techniques can guide the development of low-temperature applications. For instance, hydrogen energy has recently drawn notable attention for its significance in sustainability development yang2022status ; zhou2022green . LH2 is the main fuel of low-temperature rockets cecere2014review , and nowadays are also widely used in the industry for civil applications kumar2023synergy ; liu2021production ; tollefson2010hydrogen for their high density and low working pressure li2019review . In the condition of storage and transportation, the contact of LH2 with metallic containers may result in hydrogen embrittlement and potential risks of catastrophic fracture under loads. The packaging and reliability design of quantum circuits should also be aware of the material issues at low T brecht2017micromachined . The mechanical performance of metallic alloys at extremely low-T conditions ( \sim 20 K) is thus of concern qiu2021research . Specifically, material selection in engineering is mainly based on the match between the requirement and performance of existing materials that usually depend on the service conditions anoop2021review . For example, most materials tend to increase their strength and decrease their ductility and toughness at low T duthil2015material ; mcclintock1960mechanical . A ductile-to-brittle transition (DBT) may occur and result in limited yield, which depends on the microstructures of materials pineau2016failure . This effect is stronger in body-centered cubic (BCC) materials with less ductile performance than that of hexagonal closely-packed (HCP) and face-centered cubic (FCC) materials zhang2023unified (Fig. 1 c). Because of the diversity in the temperature ranges and refrigeration techniques of low-T applications, scientific exploration and technical reports on the mechanical properties of metallic alloys are highly heterogeneous. Collecting the research data could thus deepen our understanding of material performance for existing applications at cryogenic conditions and elucidate the potential challenges in deploying future technologies. In this work, we construct a database including tensile and impact test data reported in 715 scientific articles, which can be used as a reference for the applications of metallic alloys at low-T conditions. Methods The workflow follows our previous work on fatigue data zhang2023fatigue ; zhang2023fatigue2 , which includes content acquisition, data extraction, and database construction. The database consists of both research metadata and scientific data. Metadata includes information such as author, publication year, and keywords. Scientific data includes research data such as material types, chemical compositions, processing and test conditions, and mechanical properties. For the lack of a unified data description for low-T mechanical properties reported in figures and tables, manual extraction and correction are needed. Content acquisition After collecting the keywords of ‘low temperature’, ‘alloys’, and ‘mechanical properties’, and compiling them into the search formula (Table 1 ), Web of Science (WoS) returned 8439 article records and the metadata are obtained through the ‘export’ function. In search queries, Web of Science automatically applies the ‘stemming’ rules, which reduce word forms to their roots. The articles are classified through a natural language processing (NLP) classification model liu2019roberta according to their abstracts, the results of which ( 715 articles) are corrected through manual inspection. The documents are collected through their digital object identifiers (DOIs) in Extensible Markup Language (XML), Hyper Text Markup Language (HTML) or Portable Document Format (PDF) formats. Articles from Elsevier are downloaded through the official application programming interface (API), while articles from other publishers are automatically downloaded using open-source code article-downloader kim2017machine or manually downloaded from the publishers’ websites. XML/HTML can be automatically parsed and converted into structured text through computer codes, while PDF requires manual processing in large part. Data extraction Images in the articles are directly collected from the XML/HTML files or extracted from the PDF documents using PyMuPDF . Figures with multiple panels are automatically segmented using a rule-based code. Figures presenting mechanical properties are screened through a convolutional neural network (CNN) model (ResNet he2016deep ). Strength and fracture elongation data in images are extracted using the MATLAB code IMageEXtractor zhang2023fatigue2 . Table data in XML/HTML files are collected by using table extractor jensen2019machine and saved as a JSON format file for subsequent text mining, while table data in PDF files are extracted manually. Text in the XML/HTML and PDF documents are extracted using TEXTract zhang2023fatigue2 and PDFDataExtractor zhu2022pdfdataextractor , respectively. For literature prior to 1991, only scanned image PDFs are available, which were converted into TXT files for further text mining. Text data are mined using both GPT- 3.5 brown2020language and GLM- 4 for comparative studies. The detailed steps for conducting text mining using GPT- 3.5 are as follows. The section titles are filtered by keywords (Table 2 ) to obtain paragraphs related to the experimental methods and data. GPT- 3.5 is used to extract information on materials, processing and testing conditions, and mechanical properties from relevant paragraphs or tables. The prompts in GPT include task descriptions, examples, and the text to be processed. The task description requires GPT to

Read the complete text