A critical review on microbial degradation of petroleum-based plastics: quantitatively effects of chemical addition in cultivation media on biodegradation efficiency

Abstract

Petroleum-based plastics (PBP) with different properties have been developed to suit various needs of modern lives. Nevertheless, these well-developed properties also present the double-edged sword effect that significantly threatens the sustainability of the environment. This work focuses on the impact of microbial cultivating conditions (the elementary compositions and temperature) to provide insightful information for the process optimization of microbial degradation. The major elementary compositions in cultivation media and temperature from the literature were radically reviewed and assessed using the constructed supervised machine learning algorithm. Fifty-two literatures were collected as a training dataset to investigate the impact of major chemical elements and cultivation temperature upon PBP biodegradation. Among six singular parameters (NH₄⁺, K⁺, PO₄^3-, Mg²⁺, Ca²⁺, and temperature) and thirty corresponding binary parameters, four singular (NH₄⁺, K⁺, PO₄^3-, and Mg²⁺) and six binary parameters (NH₄⁺/K⁺, NH₄⁺/PO₄^3-, NH₄⁺/Ca²⁺, K⁺/PO₄^3-, PO₄^3-/Mg²⁺, Mg²⁺/Temp) were identified as statistically significant towards microbial degradation through analysis of variance (ANOVA). The binary effect (PO₄^3-/Mg²⁺) is found to be the most statistically significant towards the microbial degradation of PBP. The concentration range, which locates at 0.1-0.6 g/L for Mg²⁺ and 0-2.8 g/L for PO₄^3-, was identified to contribute to the maximum PBP biodegradation. Among all the investigated elements, Mg²⁺ is the only element that is statistically and significantly associated with the variations of cultivation temperature. The optimal preparation conditions within ± 20% uncertainties based upon the range of collected literature reports are recommended. Five representative cultivation elementary compositions (NH₄⁺, K⁺, PO₄^3-, Mg²⁺, and Ca²⁺) and temperature were reviewed from fifty two different literature reports to investigate their impacts on the microbial degradation of PBP using supervised machine learning algorithm. The optimal cultivation conditions based upon collected literature reports to achieve biodegradation over 80% were identified.

Keywords: Artificial neuron networks; Petroleum-based plastic; Quantitative assessment; Supervised machine learning.