A recent study from Central South University in China examined how to assess cobalt content in soil.
A new study from Central South University in China showcased the development of a new approach for identifying cobalt (Co) contamination in soil using advanced machine learning (ML) techniques. The findings of this study were published in the Journal of Environmental Chemical Engineering (1).
Co is a transition metal on the periodic table. It is considered an essential micronutrient for human beings and for plants (2). However, the questions that scientists are still pondering is how much Co is good for plants before it becomes toxic (3). There is a need for a nondestructive, inexpensive, and rapid method to help detect Co contamination in plants.
Gardener is digging soil with a shovel at spring green outdoors background. | Image Credit: © vitanovski - stock.adobe.com
Historically, the identification of Co contamination has been limited to small-scale studies and simplistic analytical methods, leaving large areas underassessed (1). Chongchong Qi and his research team at Central South University explored a new method to determine cobalt contamination by leveraging eight different ML algorithms combined with visible and near-infrared (NIR) reflectance spectroscopy to create a robust, large-scale model capable of accurately classifying Co content in soil (1).
Qi and his team referred to data set of 18,675 topsoil samples, which they collected and analyzed to train and validate the models (1). To improve the performance of the tested models, the researchers deployed two methods: principal component analysis (PCA) and rigorous hyper-parameter tuning (1). These improvements were measured using multiple evaluation indicators to ensure the highest accuracy and reliability (1).
Among the various ML models the researchers tested, the eXtreme Gradient Boosting (XGB) algorithm performed the best, with an area under the curve (AUC) value of 0.901 on the training set and 0.904 on the testing set (1). These results indicate a high level of precision in distinguishing between contaminated and non-contaminated samples.
Following the model training and validation, the optimal XGB model was applied to a comprehensive United States soil spectral data set. This application revealed several states, including Utah, Arizona, New Mexico, North Dakota, Arkansas, Mississippi, and Alabama, as having a higher risk of Co contamination (1). These findings are crucial for environmental managers and policymakers because they highlight specific regions where intervention and remediation efforts may be urgently needed (1).
This study also lays the groundwork for future research in this space. The methodology and framework developed by Qi’s team can be adapted for the detection and assessment of other hazardous elements in soil, thereby broadening the scope of environmental protection efforts (1). This study exemplifies how cutting-edge technology can be harnessed to address complex environmental challenges, ultimately contributing to safer and healthier ecosystems (1).
As cobalt is increasingly used in various industrial applications, including rechargeable batteries and aerospace components, understanding its environmental impact becomes ever more important. The ability to accurately detect and map cobalt contamination on a large scale ensures that potential risks to human health and the environment can be managed proactively (1). By combining machine learning algorithms with spectral analysis, they have developed a powerful tool for identifying and managing cobalt contamination in soil.
(1) Zhou, N.; Hu, T.; Wu, M.; et al. Comparative analysis of machine learning algorithms for identifying cobalt contamination in soil using spectroscopy. J. Environ. Chem. Eng. 2024, 12 (5), 113328. DOI: 10.1016/j.jece.2024.113328
(2) Hu, X.; Wei, X.; Ling, J.; Chen, J. Cobalt: An Essential Micronutrient for Plant Growth? Front Plant Sci. 2021, 12, 768523. DOI: 10.3389/fpls.2021.768523
(3) Srivastava, P.; Bolan, N.; Casagrande, V.; et al. In Appraisal of Metal(loids) in the Ecosystem, Chapter 5 - Cobalt in soils: sources, fate, bioavailability, plant uptake, remediation, and management. 2022, 81–104. DOI: 10.1016/B978-0-323-85621-8.00007-8
AI and Dual-Sensor Spectroscopy Supercharge Antibiotic Fermentation
June 30th 2025Researchers from Chinese universities have developed an AI-powered platform that combines near-infrared (NIR) and Raman spectroscopy for real-time monitoring and control of antibiotic production, boosting efficiency by over 30%.
Toward a Generalizable Model of Diffuse Reflectance in Particulate Systems
June 30th 2025This tutorial examines the modeling of diffuse reflectance (DR) in complex particulate samples, such as powders and granular solids. Traditional theoretical frameworks like empirical absorbance, Kubelka-Munk, radiative transfer theory (RTT), and the Hapke model are presented in standard and matrix notation where applicable. Their advantages and limitations are highlighted, particularly for heterogeneous particle size distributions and real-world variations in the optical properties of particulate samples. Hybrid and emerging computational strategies, including Monte Carlo methods, full-wave numerical solvers, and machine learning (ML) models, are evaluated for their potential to produce more generalizable prediction models.
Combining AI and NIR Spectroscopy to Predict Resistant Starch (RS) Content in Rice
June 24th 2025A new study published in the journal Food Chemistry by lead authors Qian Zhao and Jun Huang from Zhejiang University of Science and Technology unveil a new data-driven framework for predicting resistant starch content in rice
New Spectroscopy Methods Target Counterfeit Oral Medication Syrups
June 23rd 2025Researchers at Georgia College and Purdue University have developed a fast, low-cost method using Raman and UV–visible spectroscopy combined with chemometric modeling to accurately screen and quantify active ingredients in over-the-counter oral syrups, helping to fight counterfeit medications.
Short Tutorial: Complex-Valued Chemometrics for Composition Analysis
June 16th 2025In this tutorial, Thomas G. Mayerhöfer and Jürgen Popp introduce complex-valued chemometrics as a more physically grounded alternative to traditional intensity-based spectroscopy measurement methods. By incorporating both the real and imaginary parts of the complex refractive index of a sample, this approach preserves phase information and improves linearity with sample analyte concentration. The result is more robust and interpretable multivariate models, especially in systems affected by nonlinear effects or strong solvent and analyte interactions.