摘要
利用重量法精确配制不同肼质量分数的肼-70样品集,采用偏最小二乘法建立了肼质量分数的近红外光谱模型,以快速测定肼-70的纯度。样品恒温时间为5 min,光谱最佳预处理方式:均值中心化,一阶导数,21点平滑,模型最佳主因子数为4。近红外光谱法与碘酸钾直接滴定法测定结果相比相对偏差绝对值小于0.13%,经t检验,两种方法测定结果无显著性差异。近红外光谱法快速、准确,可用于肼-70产品的质量控制。
A series of hydrazine-70 sample sets of different content of hydrazine were prepared accurately by accurate weight method. The calibration model of hydrazine content was established by use of partial least squares regression for rapid determination of hydrazine-70 purity. The optimized stabilization time was 5 min, and the best spectral pretreatment methods were mean centre, first derivative and 21 point smoothing. The optimal principal factor number of spectrum model was 4. Compared with titration method, the relative deviation absolute value of determination results of near infrared spectroscopy was less than 0.13%. It was indicated that these two methods had no significant difference by t test. The near infrared spectroscopy method is rapid and accurate, and can be applied to the quality control of hydrazine-70.
出处
《化学分析计量》
CAS
2014年第5期64-67,共4页
Chemical Analysis And Meterage
关键词
肼-70
近红外光谱
偏最小二乘法
near-infrared spectroscopy
partial least squares regression