Monitoring pilot-scale polyhydroxyalkanoate production from fruit pulp waste using near-infrared spectroscopy

Research output: Contribution to journalArticlepeer-review

7 Citations (Scopus)


Given the current plastic disposal global crisis, biodegradable polymers may have an important role in leading the way towards a more sustainable plastic industry. Polyhydroxyalkanoates (PHA) are biobased and biodegradable aliphatic polyesters synthesized by bacteria and stored as intracellular granules. In this study, PHA was produced at pilot scale by a mixed microbial culture (MMC) fed with fermented fruit pulp waste. The PHA accumulation reactor was monitored at line by near-infrared (NIR) spectroscopy to develop and optimize partial least squares (PLS) calibration models for predicting bulk PHA concentration and intracellular PHA content. The PLS models were subjected to internal cross-validation (62 spectra) and the ones with best performance were validated with an external test set (12 spectra). For bulk PHA concentration, a root mean squared error of prediction (RMSEP) of 0.69 g/L and a coefficient of determination (R2) of 0.89 were attained, while intracellular PHA content was predicted with a RMSEP of 14.6% and a R2 of 0.86. These NIR-based calibration models demonstrated a great potential for the real-time monitoring of the MMC pilot-scale PHA production from a complex fruit pulp waste substrate. This approach can be used for in situ control of the pulse-wise feeding strategy of the PHA accumulation stage, minimizing PHA consumption triggered by carbon source depletion and contributing for an improved global process efficiency and productivity.

Original languageEnglish
Article number108210
JournalBiochemical Engineering Journal
Publication statusPublished - Dec 2021


  • Bioreactor monitoring
  • Fruit pulp waste
  • Mixed microbial culture
  • Near-infrared spectroscopy
  • Polyhydroxyalkanoates


Dive into the research topics of 'Monitoring pilot-scale polyhydroxyalkanoate production from fruit pulp waste using near-infrared spectroscopy'. Together they form a unique fingerprint.

Cite this