Dataset Documentation

Dataset: Time-Resolved Performance and Stability of Gas Recombination Catalysts in PEM Water Electrolysis

Overview

This dataset provides time-resolved electrochemical and gas analysis data under high-pressure conditions (30 bar absolute) from long-term (approximately 1500 hours) single-cell polymer electrolyte membrane water electrolysis (PEMWE) experiments.

To facilitate direct use of the dataset, an accompanying Jupyter Notebook is provided, offering example workflows for data loading, filtering of measurement phases, and basic exploratory analysis.

A key feature of this dataset is the combined investigation of gas recombination catalysts (GRC) and hydrogen crossover besides the electrochemical data, enabled by time-resolved gas chromatography measurements of hydrogen in oxygen content (H₂ in O₂) making this dataset especially valuable for studying degradation mechanisms, gas crossover behavior, and gas recombination catalyst performance.

Four independent experiments are included:

• Reference (no gas recombination catalyst)

• PtNano

• PtBlack

• Pt/C

Each experiment is provided as a separate CSV file with identical structure, enabling direct comparison across samples.

The dataset is suitable for:

• electrochemical performance analysis

• degradation and stability studies

• hydrogen crossover evaluation under pressure

• analysis of gas recombination catalyst performance

• impedance-based diagnostics

• data-driven modeling and machine learning applications

---------------------------------------------------------

Experimental Setup (Short summary)

• Single-cell PEM water electrolyzer (4 cm² active area)

• Membrane electrode assembly based on Nafion N115

• Cathode pressure: 30 bar absolute

• Water temperature: 80 °C

• Long-term operation: ~1500 hours

Further details on materials, cell design, and experimental setup are provided in the accompanying documentation PDF file and the original publication.

---------------------------------------------------------

File Structure

The dataset consists of four CSV files:

• Each file corresponds to one sample (Reference, PtNano, PtBlack, Pt/C)

• All files share the same column structure

• Time is reported as elapsed time starting from the beginning of each individual experiment

General file format:

• Encoding: UTF-8

• Delimiter: comma (,)

• Decimal separator: dot (.)

---------------------------------------------------------

Data Content

The dataset combines multiple types of measurements:

• Cell performance data (current, voltage)

• Electrochemical impedance spectroscopy (EIS)

• Gas chromatography (GC) measurements of hydrogen in oxygen

• Auxiliary measurements (temperature, pressure, water conductivity)

• Measurement protocol identifiers

Due to the combination of different measurement systems, the dataset contains variables with different sampling rates.

---------------------------------------------------------

Column Description

The following columns are included in each dataset:

• elapsed_time_s
  Overall measurement time in seconds. Time starts at 0 for each dataset.

• Temp_degC
  Cell block temperature in °C (measured via Pt100 sensor).
  Note: This signal may contain occasional irregularities and measurement artifacts.

• I_mA
  Applied electrical current in mA.

• E_V
  Measured total cell voltage in V.

• freq_Hz
  Frequency of EIS measurement points in Hz.
  This column contains NaN values during non-EIS operation.

• Z_magnitude_Ohm
  Magnitude of the cell impedance in Ohm.
  Only populated during EIS measurements.

• Z_phase_deg
  Phase angle of the cell impedance in degrees.
  Only populated during EIS measurements.

• conductivity_uScm
  Conductivity of the circulating water in µS/cm.

• pressure_bar_abs
  Cathode pressure in bar absolute.

• H2_uncorrected_vol_percent
  Raw hydrogen concentration in oxygen measured by gas chromatography (in vol.%).

• H2_corrected_vol_percent
  Dilution-corrected hydrogen concentration in oxygen (in vol.%).
  This is the physically meaningful value and should be used for analysis.

• measurement_type
  Identifier of the current step in the experimental protocol. Possible values include:

  • pressure_build_up

  • conditioning

  • POL_GEIS

  • H2_char

  • CV_cycles

  • OCV

  • CstC_12A

  • GEIS_12A

  • CstC_1A

  • GEIS_1A

• measurement_type_counter
  Counter for each measurement type, enabling identification of repeated sequences
  (e.g., multiple polarization curves across the experiment).

---------------------------------------------------------

Measurement Protocol

The experimental procedure consists of repeated sequences of conditioning, characterization, and stress phases.

General structure:

1. Pressure build-up phase (ID: pressure_build_up)
   Stepwise current holds are applied to reach the target cathode pressure (30 bar absolute).

2. Conditioning phase (ID: conditioning)
   Long-term operation at defined current and voltage to stabilize the system.

3. Characterization phase
   Repeated characterization including:

   • polarization curves with galvanostatic EIS measurements (ID: POL_GEIS)

   • hydrogen crossover measurements under diferent constant currents (ID: H2_char)

   • another polarization curves with galvanostatic EIS measurements (ID: POL_GEIS)

4. Stress phase
   Accelerated degradation protocol including:

   • voltage cycling (ID: CV_cycles)

   • open circuit voltage (OCV) measurements (ID: OCV)

   • long constant-current holds at 12A high followed by a GEIS (ID: CstC_12A, GEIS_12A)

   • long constant-current holds at 1A high followed by a GEIS (ID: CstC_1A, GEIS_1A)

These characterization and stress phases are repeated multiple times over the total duration of approximately 1500 hours.

A detailed schematic of the protocol is provided in the accompanying PDF documentation.

---------------------------------------------------------

Sampling Rates and Data Density

• Electrical signals (current, voltage, temperature, pressure, conductivity) are recorded continuously with high temporal resolution (every second).

• Gas chromatography (GC) measurements are performed at significantly lower frequency (typically every 1.5 minutes).

• EIS data are recorded only during dedicated measurement steps.

As a result:

• Many columns (especially gas analysis and EIS data) contain NaN values outside their respective measurement phases.

• This is expected behavior and reflects the experimental design.

---------------------------------------------------------

Missing Data and Known Issues

• NaN values indicate:

  • parameters not measured at a given time (e.g., EIS outside EIS phases)

  • lower sampling rates (e.g., gas chromatography)

  • temporary measurement interruptions

• Between approximately 300 h and 500 h:

  • pressure and conductivity data are partially unavailable due to a temporary connection error for the PtBlack and PtNano dataset

• Temperature measurements:

  • represent the cell block temperature, not the exact reaction temperature

  • may show deviations due to self-heating at high current densities

---------------------------------------------------------

Data Usage Notes

• For hydrogen crossover analysis, use H2_corrected_vol_percent only.

• EIS data should be filtered using:

  • freq_Hz not NaN → EIS data

• Constant-current segments can be isolated using:

  • freq_Hz = NaN

• The measurement_type and measurement_type_counter columns are essential for:

  • extracting specific protocol steps

  • comparing repeated measurements across time

---------------------------------------------------------

Additional Files

The dataset is accompanied by:

• Detailed dataset documentation (PDF)
  Includes full experimental description, protocol diagrams, and extended explanations.

• Jupyter Notebook
  Provides example code for:

  • loading the dataset

  • filtering specific measurement phases

  • basic exploratory analysis

---------------------------------------------------------

License

This dataset is licensed under the Creative Commons Attribution 4.0 International (CC BY 4.0).

Citation

If you use this dataset, please cite the original publication linked to this dataset: https://doi.org/10.1039/d6ey00106h

Notes

This dataset is designed to support both classical electrochemical analysis and modern data-driven approaches. Users are encouraged to consult the accompanying documentation and notebook for correct interpretation of the measurement protocol and data structure.