出 处:《冶金分析》2023年第10期1-9,共9页Metallurgical Analysis
基 金:国家重点研发计划资助项目(No.2021YFF0700100)。
摘 要:等离子光谱是激光诱导击穿光谱技术(LIBS)的重要探测对象,其主要包含样品原子从高激发态经不同路径退激到低激发态产生的特征光谱,辨别不同特征光谱是认识测量样品特性的一种有效方式。利用“就近原则”法识别特征光谱常见于各类商用光谱测量软件中,经研究发现,“就近原则”法识别特征光谱时不可避免地存在偶然符合误差,从而导致谱线识别错误率高。实验系统分析了“就近原则”法中偶然符合误差产生的机制。激光诱导击穿光谱的特征峰与元素特征谱线的波长差直接影响谱线真实符合数,实验表明,谱线真实符合仅存在于某一极小波长差范围内,利用这一特性,实现了测量光谱的波长漂移自校正。另外,建立了弱偶然符合模型,以提高谱线识别准确度,并实现了样品元素和谱线自动化识别。弱偶然符合模型克服了“就近原则”法中依赖经验判断实验谱线与元素特征谱线的远近,以及仅依赖单条特征谱线判断元素是否存在等缺点。在弱偶然符合模型中,加入单个元素谱线符合数阈值T_(line)、波长差阈值T_(wave)和光谱寻峰强度阈值T_(peak),分析了阈值T_(line)、T_(wave)和T_(peak)对元素识别的影响,并优化阈值使偶然符合产生的误识别元素的期望个数小于1,且最大化保留谱线的真实符合。利用弱偶然符合模型,对标准样品FeNi合金(GBW(E)010081)、玻璃样品NIST610和NIST612的激光诱导击穿光谱进行了分析,识别的元素个数分别为8个、35个和18个,且元素识别的正确率达到100%;弱偶然符合模型对于发射强度较高的特征谱线也具有较高的识别率,分析结果表明光谱特征峰相对强度大于10%的谱线的识别率高于60%。Plasma spectroscopy is an important detection object of the laser-induced breakdown spectroscopy technology.It mainly includes the characteristic spectra of sample atoms from highly excited state to low excited state through different paths.The identification of different characteristic spectra is an effective way to understand the characteristics of samples.The“Adjacent Principle”method is commonly used to identify the characteristic spectra in various commercial spectral measurement software.It was found that there were inevitable accidental coincidence errors during the identification of characteristic spectra by the“Adjacent Principle”method,resulting in a high error rate of spectral line identification.The mechanism of accidental coincidence error in the“Adjacent Principle”method was systemically analyzed.The wavelength difference between the characteristic peak of the laser-induced breakdown spectroscopy and the characteristic spectral line of the element directly affected the number of true coincidences.The experimental results showed that the true coincidence of the spectral line only existed in a certain wavelength difference range.Based on this characteristic,the self-calibration of wavelength drift of measuring spectra was realized.In addition,a weak accidental coincidence model was established to improve the accuracy of spectral line identification,and the automatic recognition of sample elements and spectral lines was realized.The weak accidental coincidence model overcame the shortcomings of the"Adjacent Principle"method,such as relying on experience to judge the distance between the experimental spectral line and the element characteristic spectral line,and relying only on a single characteristic spectral line to judge whether the element exists.In the weak accidental coincidence model,the threshold of coincidence number T_(line),wavelength difference T_(wave)and spectral peak seeking intensity T_(peak)were added.The effects of thresholds T_(line),T_(wave)and T_(peak)on the element reco
关 键 词:激光诱导击穿光谱(LIBS) 谱线识别 元素识别 弱偶然符合
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...