阿莫西林和多四环素在紫外光谱区间吸收峰附近有多重共线现象,紫外分光光度法难以准确测出2种物质各自含量.本文通过L16(4^5)正交实验设计,采集200~250nm波长混合样品吸收光谱,数据经标准化处理,建立偏最小二乘回归预报数学模型.结果显示,该法计算阿莫西林和多四环素的预测相对误差(RPE)分别为3.13%和6.42%,平均加标回收率分别为100.34%和100.71%.该法简便快捷,准确可靠,样品无需预先分离.
The existence of colinear and overlapping phenomena in the ultraviolet absorption spectra of amoxicillin and doxycyeline makes it difficult to use ultraviolet spectrophotometry to determine the respective contents in a mixture of both components. L16 (45) orthogonal experimental design was applied in this study and absorption data from of 200 to 250 nm of their mixture were obtained with UV scan procedure. Then a data normalization was made and a PLS (partial least square regression) predicting model was developed by using DPS 8.01 software. It was shown that this method was accurate and reliable, with the RPE% for both amoxicillin and doxycycline were 3.13% and 6. 42% , respectively, and the average recovery rate were 100. 34% and 100.71% ,respectively. It was also simple and rapid without the necessity of separating the samples beforehand.