Quantitative structure-property relationship for predicting chlorine demand by organic molecules.

Area of Science:

• Environmental Chemistry
• Water Treatment Technologies
• Computational Chemistry

Background:

• Conventional methods for predicting chlorine demand rely on bulk parameters, neglecting molecular structure.
• Dissolved organic matter (DOM) significantly influences chlorine demand in water treatment.
• Understanding molecular reactivity is crucial for accurate disinfectant demand prediction.

Purpose of the Study:

• To develop a Quantitative Structure-Property Relationship (QSPR) model for predicting hypochlorous acid (HOCl) demand.
• To account for the structural properties of individual molecules influencing chlorine demand.
• To provide a more accurate prediction of disinfectant demand compared to traditional methods.

Main Methods:

• Developed a QSPR model for HOCl demand using eight constitutional descriptors.
• Utilized multiple linear regression for model calibration with a leave-many-out approach.
• Validated the model using calibration, cross-validation, and external datasets (N=159 and N=42).

Main Results:

• Achieved an average R² of 0.86 and a standard error of regression of 1.24 mol HOCl/mol compound during calibration.
• Demonstrated robust model performance with internal cross-validation (q² = 0.85) and external validation (q² = 0.88).
• External validation showed a root-mean-square error of prediction of 1.17 mol HOCl/mol compound.

Conclusions:

• The developed QSPR model provides a robust and accurate method for predicting chlorine demand based on molecular structure.
• The model's predictions for NOM structures align with experimental measurements, indicating practical applicability in water treatment.
• This approach offers a significant advancement over conventional methods by incorporating molecular insights into disinfectant demand prediction.