火花放电原子发射光谱仪稳定性探讨

火花放电原子发射光谱仪在国内钢铁企业大量配置。设备按需求配置待测元素的原始校准工作曲线,用户在现场进行全局标准化、曲线确认后,仪器即可投入分析工作。其间,选取适宜的监控样品对测量结果进行定期监测,而监控间隔取决于仪器的稳定性。稳定性是监控测量结果的准确度,即精密度和正确度。对火花放电原子发射光谱仪进行稳定性评定的方案设计,将目标测量时间分为多个时间节点,选取合适的光谱标样,执行标准GB/T 4336—2016,在每个节点对标准样品进行测量。利用标准中给出的精密度数据,采用精密度、正确度判据及χ²统计量对测量数据进行以下5项指标检验:各时段内数据重复性(y_i1-y_i2)、各时段内数据正确度(公式)、时段间重复性(s²_rt)、时段间总精密度(公式)及总均值正确度(-μ₀),各指标顺次逐步评价,前二者评价时,以保留最多组连续数据为准;后三者评价时,从最后数据开始剔除至满足相关检验要求。整理剔除后的最终数据,计算各元素连续稳定的最短时间,即为该元素在该含量水平下的稳定性时间上限T_MAX,以所有元素所有水平下的T_MAX作为仪器的稳定性时间上限。论文所用光谱仪稳定性时间上限为6 h,在此时间内,仪器无需任何校正。采用T_MAX指标时实验室不需要监控实时测试数据,采用该指标评价火花放电原子发射光谱仪能够为实验室节省时间和成本。

Discussion on stability of spark discharge atomic emission spectrometer

Spark discharge atomic emission spectrometer (Spark-AES) are equipped in many steel and iron enterprises in China. The original calibration curves of testing elements are prepared in the equipment in response to the needs of the user. After standardization and verification by the users on site, the instrument can be used for analysis. In the meantime, the appropriate verification samples are selected to monitor the measurement results at regular intervals. The monitoring interval usually depends on the stability of the instrument. Stability means the accuracy of the measurement results, i.e., the precision and trueness. The stability assessment methods for Spark-AES were designed. The target measurement time was divided into several time nodes. The proper standard sample for spectral analysis were selected and measured at each time node according to standard GB/T 4336-2016. Using the precision data given in the standard, the following five indicators were tested by precision, trueness criteria and χ² statistics: the precision of the measurement data in each period (y_i1-y_i2), the trueness of the measurement data in each period (_i-μ₀), the repeatability of each period (s²_rt), the total precision of each period (s²_i) and the accuracy of the total mean value (-μ₀). The criteria were evaluated step by step. During the evaluation of the first two indicators, the maximum group of consecutive data was retained. During the evaluation of the latter indicators, the final data were removed until the relevant test requirements were met. The minimum continuous stable time of each element was calculated according to the final data after collation, i.e., the upper limit of the stability time (T_MAX) for the element at such content level. T_MAX at all levels for all elements was used as the instrument's maximum stability time. The T_MAX of the spectrometer used in the paper was 6 h. Any calculation was required for the instrument within this period. The monitoring of real-time test data was not required when the T_MAX indicator was adopted in laboratory. The use of index to evaluate Spark-AES could save time and cost for the laboratory.