近红外漫反射光谱法定性定量分析头孢拉定胶囊.docVIP

近红外漫反射光谱法定量分析头孢拉定胶囊 潘颖1 沈漪1 项竞佐1 刘全2 1上海市药品检验所抗生素室， 上海200233 2 美国Thermo Nicolet仪器公司上海代表处，上海200031 摘要 目的：采用近红外（NIR）漫反射光谱法对不同生产厂家的头孢拉定胶囊进行快速定量分析。方法：按头孢拉定胶囊配方组成配制含主药头孢拉定浓度范围从5.01% ~ 91.24%的30个实验室样品，并收集来源于7个厂家的49批工业样品，采集其NIR光谱。采用偏最小二乘回归法建立NIR光谱信息与样品组成间的定量分析模型，将其用于对验证样品进行预测分析，并对该方法的加样回收率进行考察。结果：定量分析模型对21验证样品的的预测均方差RMSEP为1.35%，预测值与真值的相关系数R为0.9968，加样平均回收率为99.7%，RSD为0.7%（n=6）。结论：用近红外漫反射光谱法对头孢拉定胶囊进行定量分析的方法简便快速，结果准确可靠，可推广用于工业现场的实时在线检测。 关键词 近红外漫反射光谱法 头孢拉定胶囊 偏最小二乘回归 定量分析 Quantitative Analysis of Cefradine Capsule by Near-infrared Spectroscopy Pan Ying1, Shen Yi1, Xiang Jingzuo1, Liu Quan2 Shanghai Institute for Drug Control, Shanghai 200233) Thermo Nicolet Corporation Shanghai Representative Office, Shanghai 200031) Abstract Objective To determine the principal component in cefradine capsule by near-infrared(NIR) spectroscopy. Method 30 samples with a concentration range from 5.01% to 91.24% for cefradine were prepared in laboratory and 49 industry samples from 7 different factories were collected. Then their NIR spectra were collected on an Antaris FT-NIR analyzer. Partial least square regression (PLSR) algorithm was used to build the calibration model between the concentration of 59 calibration samples and their NIR spectra. At last, the model was validated by the other 20 validation samples. The recovery of the method has also been examined. Result The root mean square error of prediction (RMSEP) was 1.35%. The correlation coefficient between the true value and prediction value was 0.9968. The average recovery was 99.7% (RSD 0.67%, n=6). Conclusion The results show that the presented method is convenient, fast and has a good performance and can be applied to on-line detection in industrial process. Key words Near -Infrared Spectroscopy; Cefradine capsules; Partial least square regression (PLSR); Quantitative analysis 近红外（near-infrared，NIR）光谱技术是一种快速、无损分析方法，其光谱波长范围是780～2500nm（11000～4000cm-1），主要谱峰为有机物分子C-H、N-H和O-H等含氢基团的倍频与合频振动吸收所产生，光谱特性稳定，非常适合于药品原辅料及制剂的定性定量分析[1-5]