荧光相关光谱测定溶液中胭脂红的含量

发布时间：2019-03-05 20:36

【摘要】：使用FLS920P型荧光光谱仪测量了20个合成色素胭脂红溶液样本的荧光发射谱,实验表明:胭脂红的最佳激发波长为300nm,在此波长激发光下,荧光峰值波长为440nm。同时测量相同条件下超纯水的光谱数据作为参考光谱,进行与胭脂红溶液光谱数据的相关计算,构建以浓度为外扰的荧光相关光谱。采用sym8小波函数4尺度降噪,将降噪后的同步相关光谱数据、自相关光谱数据应用偏最小二乘回归(PLSR)算法进行预测,建立溶液中胭脂红含量的定量模型,结果表明:采用同步相关光谱建模的预测相关系数为99.863%,预测均方根误差为0.414μg·mL~(-1);而采用自相关光谱建模的预测相关系数为99.940%,预测均方根误差为0.303μg·mL~(-1)。对比可知,自相关光谱数据有效地避免了信息冗余,预测结果更为可靠。该方法无需样本处理,操作简单,为食品安全检测提供了一种新的思路。
[Abstract]:The fluorescence emission spectra of 20 samples of synthetic pigment rouge solution were measured by FLS920P fluorescence spectrometer. The results showed that the optimum excitation wavelength of carmine was 300 nm, and the peak fluorescence wavelength was 440 nm under the excitation light of this wavelength. At the same time, the spectral data of ultra-pure water under the same conditions were measured as reference spectra, and the correlation between the spectral data of rouge solution and the spectral data of rouge solution was calculated, and the fluorescence correlation spectrum with concentration as external disturbance was constructed. Using sym8 wavelet function 4 scale de-noising, the simultaneous correlation spectral data and autocorrelation spectral data after denoising are predicted by partial least squares regression (PLSR) algorithm, and the quantitative model of rouge content in solution is established. The results show that the prediction correlation coefficient is 99.863% and the root mean square error is 0.414 渭 g 路mL ~ (- 1). The prediction correlation coefficient is 99.940% and the root mean square error is 0.303 渭 g 路mL ~ (- 1). The comparison shows that the autocorrelation spectral data can effectively avoid the redundancy of information, and the prediction results are more reliable. This method does not require sample processing and is simple to operate, which provides a new idea for food safety detection.
【作者单位】：江南大学理学院;江苏省轻工光电工程技术研究中心;
【基金】：国家自然科学基金项目(61178032) 中央高校基本科研业务费专项资金项目(JUSRP51517,JUSRP51628B)资助
【分类号】：O657.3;TS201.6

【相似文献】