基于奇异值分解的光子相关光谱滤波方法研究

Study on the filtering algorithm of photon correlation spectroscopy based on singular value decomposition

针对光子相关光谱颗粒测量法在测量超细纳米颗粒时,容易受噪声影响,导致拟合误差较大的问题,提出了一种基于奇异值分解的光子相关光谱滤波方法。其处理步骤为:利用颗粒系的光强自相关函数数据构造Hankel矩阵H;对矩阵进行奇异值分解;根据奇异值的大小分布,确定噪声级别和重建参数r;从重建矩阵H1中提取经滤波后的光强自相数据,再通过传统方法进行拟合,得到颗粒的粒径分布。实验中采用30nm标准乳胶球单分散颗粒系,以及30nm和100nm标准乳胶球双分散颗粒系进行实验对比。结果证明:基于奇异值分解的光子相关光谱滤波法有效地提高了测量准确性。

During the measurement of the ultrafine nanoparticles＇s size with the method of PCS（Photon Correlation Spectroscopy）, the fitting results are always susceptible to the noise and have quite large errors. To solve the problem, the filtering algorithm of photon correlation spectroscopy based on singular value decomposition is proposed. Its proce- dure is： the Hankel matrix H with the intensity autocorrelation data is constructed; the singular value decomposition of H is calculated; the reconstruction parameters r with the singular value of H1 are deter- mined; the filtered light intensity autocorrelation data from the reconstruction matrix H1 are extracted, and fitting with the traditional method, and then the particle size distribution is obtained. The experiment is carried out in two different particle dispersions, one is 30nm standard monodisperse la- tex particle dispersion, and another is 30nm and 100nm standard double dispersion latex particle dis- persion. The results show that, the filtering algorithm of photon correlation spectroscopy based on singular value decomposition can improve the measurement accuracy effectively.