AIMD-L: An automated laboratory for high-throughput characterization of structural materials for extreme environments

TL;DR

The Artificial Intelligence in Materials Design Laboratory (AIMD-L), an automated, high-throughput facility for characterizing the microstructure and properties of structural metals and ceramics, with an emphasis on materials in extreme environments is presented.

Abstract

Rapid developments in artificial intelligence and machine learning as applied to materials science are creating an urgent need for experimental data, which can be provided by high-throughput and autonomous laboratories. To date most demonstrations of such laboratories have focused on functional materials, with less attention paid to structural materials. We present here the Artificial Intelligence in Materials Design Laboratory (AIMD-L), an automated, high-throughput facility for characterizing the microstructure and properties of structural metals and ceramics, with an emphasis on materials in extreme environments. AIMD-L has two custom instruments for characterization of structural materials: HELIX for shock studies of materials, and MAXIMA for X-ray diffraction and X-ray fluorescence spectroscopy. Specifically designed for high-throughput studies, HELIX and MAXIMA are each capable of collecting data at rates two to three orders of magnitude faster than conventional systems. A third experimental station, SPHINX, is a commercial nanoindenter modified for integration into the automated workflow of AIMD-L. A user (which may be human or an AI agent) directs the experiments to be carried out by means of a centralized control program. The experimental stations are linked by a conveyance that moves samples around the lab, with a robot at each station for sample transfer in/out of the instrument. The experimental stations also communicate with a common data layer that streams data autonomously from each instrument to a data portal, where their arrival triggers automated workflows for data reduction and analysis. The processed data are immediately available to the human operator or agentic AI, forming a closed loop for rapid decision-making and experimental control.

Figures (5)

• Figure 1: The Artificial Intelligence for Materials Design Laboratory (AIMD-L) at Johns Hopkins University comprises five experimental stations surrounding a central conveyance and robotic sample handling system (Sec. \ref{['sec:robotics']}). The experimental stations include X-ray diffraction and X-ray fluorescence (MAXIMA, Sec.~\ref{['sec:maxima']}), laser microflyer impact for shock studies (HELIX, Sec. \ref{['sec:helix']}), and nanoindentation (SPHINX, Sec. \ref{['sec:nanoindentation']}). Profilometry is available for characterization of surface topography, and the flex station allows for incorporation of new capabilities.
• Figure 2: Standard holder for samples in AIMD-L. The holder has a $\qty{34}{mm}\times \qty{34}{mm}$ square through-hole, allowing access to both front and back surfaces for experiments. Samples are held securely by either adhesive or tape. The large flat area to the right of the square hole is to provide a space for the robotic vacuum grippers to pick up the holder.
• Figure 3: Control signals and data flow in AIMD-L. A human or AI agent provides instructions to the run manager through an API (either directly or, for the human, through a web interface). The run manager directs the flow of samples through the lab and issues instructions to the various instruments on the experiments to be run. Data and metadata are streamed autonomously off the instruments to a data portal, where their arrival can trigger event-driven workflows such as automated data processing. The data and metadata are immediately available to the agent (human or AI) to enable decisions to be made for the next cycle of experiments.
• Figure 4: Overview of fabrication of compositionally-graded Cu--Ti alloy foils. (a) Sputter deposition of Cu--Ti foils onto brass substrates mounted on a carousel that rotated them past Cu and Ti sources in turn, producing nanolaminate specimens. The inset shows the shield on the Ti target, which causes a vertical concentration gradient from 0.1at.% Ti at the bottom of the substrates to 7at.% Ti at the top. (b) Following deposition the samples were removed from the brass substrates, sandwiched between alumina plates, placed in a holder, and suspended by a Nichrome wire for annealing (solutionizing) in a vacuum furnace. When annealing was complete the package was dropped directly into an oil quench bath.
• Figure 5: Example data from the three primary AIMD-L experimental stations from combinatorial Cu-Ti samples: elastic modulus and hardness from nanoindentation measurements on SPHINX, lattice parameter of Cu from MAXIMA, and Hugoniot elastic limit (HEL) from HELIX. The results are plotted as functions of composition measured via XRF in MAXIMA. Red open circles indicate regions where where is clear evidence for the presence of Cu$_4$Ti intermetallic precipitates from the XRD measurements; black filled circles indicate regions without clear evidence for precipitates.