Abstract In this paper, the proposed method has been introduced for qualitative and quantitative determination of micropore volume in nano ZSM-5 zeolite samples based on diffuse reflectance infrared Fourier transform (DRIFT) spectroscopy along with chemometric methods. For this purpose, a total of 72 samples were provided and DR-FTMIR spectra corresponding to the samples were obtained in the range of 400–4000 cm−1. In order to estimate total surface area from a nitrogen adsorption isotherm, Brauner Emmett Teller (BET) method as a reference method was used. The results are displayed that the significant adsorption peak of the mid-IR spectroscopy in 3660 cm−1 is corresponded to the Si OH groups of the zeolite external surface that is a criterion for the intensity of the mesoporosity. The intensity of the 3745 cm−1 peak is grown up with enhancing the external surface area. These changes were evaluated by non-negative matrix factorization -alternating least squares (NMF –ALS) chemometrics approach for estimating the microporosity in nano ZSM-5 zeolite samples. Back-propagation artificial neural network (BP-ANN) as a nonlinear multivariate regression technique was also used to determine the micropore volume of zeolites based on DRIFT spectroscopy. For BP-ANN model in these samples, R and minimum mean squared error (MSE) values of the testing set data with 5 hidden nodes were 9.95 × 10−1 and 4.11 × 10−7, respectively. The achieved results are presented that DRIFT spectroscopy coupled with chemometrics approaches is a precise, explicit and easy technique to analyze the volume of the micropore in ZSM-5 nano-catalyst samples.

中文翻译：

---

使用反向传播人工神经网络 (BP-ANN) 和非负矩阵分解-交替最小二乘法 (NMF-ALS) 作为化学计量学，通过漫反射红外傅里叶变换 (DRIFT) 光谱法测定纳米 ZSM-5 沸石样品的微孔体积方法

摘要 本文介绍了基于漫反射红外傅里叶变换 (DRIFT) 光谱和化学计量学方法对纳米 ZSM-5 沸石样品中的微孔体积进行定性和定量测定的方法。为此，共提供了 72 个样品，并在 400-4000 cm-1 范围内获得了与样品对应的 DR-FTMIR 光谱。为了从氮吸附等温线估计总表面积，使用 Brauner Emmett Teller (BET) 方法作为参考方法。结果表明，中红外光谱在 3660 cm-1 处的显着吸附峰对应于沸石外表面的 Si OH 基团，这是中孔强度的标准。3745 cm-1 峰的强度随着外表面积的增加而增加。这些变化是通过非负矩阵分解-交替最小二乘法 (NMF –ALS) 化学计量学方法评估的，用于估计纳米 ZSM-5 沸石样品中的微孔率。反向传播人工神经网络（BP-ANN）作为一种非线性多元回归技术也被用于基于漂移光谱确定沸石的微孔体积。对于这些样本中的 BP-ANN 模型，具有 5 个隐藏节点的测试集数据的 R 和最小均方误差 (MSE) 值分别为 9.95 × 10−1 和 4.11 × 10−7。所取得的结果表明，漂移光谱与化学计量学方法相结合是一种精确、明确且简单的技术，可用于分析 ZSM-5 纳米催化剂样品中的微孔体积。