Predicting a Wide Range of Fractal Dimensions of Salt-Induced Aggregates in Water Using a Random Forest Model

Area of Science:

• Environmental Science
• Colloid Science
• Geochemistry

Background:

• Salt-induced colloidal aggregates impact contaminant fate and transport.
• Aggregate fractal dimensions (d_f) vary with system parameters, but relationships are not fully understood.
• Predicting the full range of d_f requires further exploration.

Purpose of the Study:

• To develop a predictive model for colloidal aggregate fractal dimensions (d_f).
• To identify key physical and chemical parameters influencing d_f.
• To explore the quantitative relationship between system variables and d_f.

Main Methods:

• Development of a random forest model.
• Input variables: ionic strength, relative aggregate hydrodynamic radius, particle concentration, and primary particle radius.
• Model training and testing on diverse colloid sizes (nano to micro).

Main Results:

• The model accurately predicts the complete range of aggregate d_f.
• Ionic strength is the most influential parameter for micro-sized particles.
• For mixed nano- and micro-sized particles, ionic strength and relative aggregate size are significant.
• All input parameters showed a negative correlation with predicted d_f.

Conclusions:

• A robust random forest model can predict aggregate fractal dimensions using four simple parameters.
• The findings provide a quantitative understanding of factors controlling aggregate structure.
• This model aids in predicting contaminant transport in various environmental and engineered systems.