Abstract This study aimed to use the Attenuated Total Reflectance-Fourier transform mid-infrared (ATR-FTIR) spectroscopy coupled with chemometrics, as an alternative method for the qualitative-quantitative determinations of flocculent quality. To the best of our knowledge, there is no work that has treated the properties of flocculants by direct spectroscopic methods. The quality of the flocculent was ensured by the determination of two parameters namely, sedimentation rate (VS) and solid rate (TS), during the coagulation/flocculation process. Firstly, spectral and reference data of 36 samples from three different lots of flocculent, were analyzed by principal component analysis (PCA), Partial Least Squares Regression (PLSR) and results were used to establish calibration models. PLSR models revealed the best calibration models for predicting VS and TS values for studied samples, with the coefficient of determination (R2) of 0.99. In addition, cluster analysis (CA) and linear discriminant analysis (LDA) were adopted as differentiation methods of studied samples. On the basis of CA, three distinctive clusters were recognized. The model built by LDA could classify the samples with an accuracy value of 100%. Hence, it can be concluded that ATR-FTIR spectroscopy combined multivariate analysis can be used for the rapid prediction of flocculent quality.

中文翻译：

---

使用傅里叶变换中红外光谱和多元分析快速控制工业絮凝剂的质量

摘要 本研究旨在使用衰减全反射-傅里叶变换中红外 (ATR-FTIR) 光谱结合化学计量学，作为定性定量测定絮凝物质量的替代方法。据我们所知，目前还没有通过直接光谱方法处理絮凝剂特性的工作。在混凝/絮凝过程中，通过测定沉降率（VS）和固体率（TS）这两个参数来确保絮凝剂的质量。首先，通过主成分分析（PCA）、偏最小二乘回归（PLSR）对来自三个不同批次的絮凝剂的36个样品的光谱和参考数据进行分析，并将结果用于建立校准模型。PLSR 模型揭示了用于预测研究样品的 VS 和 TS 值的最佳校准模型，决定系数 (R2) 为 0.99。此外，聚类分析（CA）和线性判别分析（LDA）被用作研究样本的区分方法。在 CA 的基础上，识别出三个不同的集群。LDA 建立的模型可以对样本进行分类，准确率为 100%。因此，可以得出结论，ATR-FTIR 光谱联合多变量分析可用于絮凝质量的快速预测。LDA 建立的模型可以对样本进行分类，准确率为 100%。因此，可以得出结论，ATR-FTIR 光谱联合多变量分析可用于絮凝质量的快速预测。LDA 建立的模型可以对样本进行分类，准确率为 100%。因此，可以得出结论，ATR-FTIR 光谱联合多变量分析可用于絮凝质量的快速预测。