Bayesian Optimized Design for O-band Bottom-up Grown Microring Lasers

Datasets for "Bayesian Optimized Design for O-band Bottom-up Grown Microring Lasers"

These datasets support the multi-objective Bayesian optimization (MOBO) of bottom-up grown InP/InAsP multi-quantum well (MQW) microring lasers. The optimization focuses on three key objectives: minimizing the lasing threshold, targeting O-band lasing wavelengths, and maximizing yield. They include microring growth and geometry parameters, performance metrics, and power-dependent photoluminescence (PL) spectroscopy data. The datasets are organized into three components: (1) the training dataset, (2) the optimized and Design of Experiments (DoE) dataset, and (3) the remeasurement dataset.

1. Model Training Dataset

File: 'Training_Dataset.pkl'
This dataset was used to train the surrogate model used in the optimization process.

Contents:

Growth Parameters:

• Number of quantum wells
• Quantum well growth temperature
• As/P ratio (estimated from precursor flow rates in the vapor phase)
• V/III ratio during InP barrier growth
• Capping layer growth duration

Geometry Parameters:

• MQW Microring diameter
• Pitch (center-to-center spacing between MQW microrings)

Performance Metrics:

• Median lasing threshold across a field (a group of nominally identical MQW microrings)
• Median lasing wavelength across a field
• Yield: Ratio of lasing MQW microrings to total MQW microrings in the field

Associated Data:

• Lasing threshold and wavelength values were derived from power-dependent PL spectroscopy.
• Power-dependent PL spectroscopy data: 'Training_pdep.h5'
• Metadata (e.g., experimental conditions, instrument settings): 'settings_training.h5'

2. Optimized and DoE Dataset

File: 'Opt_&_DoE_Dataset.pkl'
This dataset contains measurements from MQW microring structures designed through two approaches: MOBO and a structured DoE strategy.

Contents:

Growth and Geometry Parameters:

• Same structure and format as in the training dataset

Performance Metrics:

• Identical to those in the training dataset.

Associated Data:

• Power-dependent PL spectroscopy data: 'Opt_&_DoE_pdep.h5'
• Metadata (e.g., experimental conditions): 'settings_Opt_DoE_remeasurements.h5'

3. Remeasurement Dataset

This dataset includes remeasurements of the best-performing MQW microring sample from the training dataset. These remeasurements serve to validate the consistency and reproducibility of lasing threshold and wavelength performance under updated experimental conditions.

Files:

• 'remeasured_pdep.h5': Power-dependent PL spectroscopy data
• 'remeasured_sample_D.pkl': Summary of lasing threshold, wavelength and yield
• 'settings_Opt_DoE_remeasurements.h5': Metadata
  

Simultaneous improvements in all three performance metrics were observed in the optimized MQW microring lasers, validating the effectiveness of the MOBO framework.