Skip to main navigation Skip to search Skip to main content

A machine learning-assisted engineering of activation-free hierarchical and microporous biochars for selective adsorption

  • Hamad bin Khalifa University

Research output: Contribution to journalArticlepeer-review

Abstract

Biochar surface engineering for targeted applications is critical but typically relies on environmentally and economically burdensome activation processes. Here, we present a systematic, side-by-side comparison of cellulose (C) and cellulose acetate (CA) pyrolyzed under identical, activation-free conditions. We demonstrate how a single chemical modification dictates divergent structural evolution, supported by a machine learning-assisted exploratory analysis that quantified the substrate-specific divergence in thermochemical behaviour. CA-biochars developed hierarchical structures, reaching a peak surface area of 626 m(2)/g (CA-600-2hr) and mesopore/micropore volumes of 0.26/0.13 cc/g (CA-500-1hr). Conversely, C-biochars were predominantly microporous, with C-600-2hr achieving peak micropore volume (0.21 cc/g), mesopore volume (0.030 cc/g), and surface area (590 m(2)/g), closely aligning with predicted cellulose carbonization trajectories (R-2 = 0.80, RMSE = 5.8) via a voting regressor ensemble algorithm. To verify the effectiveness of this approach, four biochars were tested in size-selective micropollutant adsorption and small-molecule CO2 capture. For planar Methylene Blue, the highly graphitic, microporous C-600-2hr had superior adsorption capacity (q(m) = 38.17 mg/g) and binding affinity (K-L = 1.96 L/mg), attributable to its compatible pore dimensions (similar to 1.9 nm) and dominant pi-pi interactions. Conversely, hierarchical CA-600-2hr excelled for bulkier Rhodamine B (q(m) = 47.62 mg/g), mitigating diffusion limitations. Dubinin-Radushkevich modelling further confirmed that C-600-2hr's superior CO2 capture (2.07 mmol/g) was driven by its optimized micropore capacity (W-0: 5.24 mmol/g) rather than surface functionalization. These findings suggest that precursor selection and pyrolysis conditions can be systematically varied to engineer biochar platforms with tunable functionalities for size-selective adsorption applications.
Original languageEnglish
Article number134905
Number of pages14
JournalBioresource Technology
Volume456
DOIs
Publication statusPublished - Sept 2026

Keywords

  • Adsorbent
  • Artificial intelligence
  • Biochar engineering
  • CO2 adsorption
  • Methylene Blue
  • Micropollutant
  • Pyrolysis
  • Rhodamine B

Fingerprint

Dive into the research topics of 'A machine learning-assisted engineering of activation-free hierarchical and microporous biochars for selective adsorption'. Together they form a unique fingerprint.

Cite this