Inductively coupled plasma-mass spectrometry (ICP-MS) was used to measure the aluminum content in the quality control sample of milk powder, and an X-control chart was established for runaway judgement and handling during the operation process. The precision and median line of the control chart for new data points obtained one year later were evaluated using Minitab software with test for equal variance and analysis of variance. The standard deviation for the previous 25 data points was 0.186 mg/kg compared to 0.126 mg/kg for the 35 new ones, while the median values of the two groups did not differ from each other. More reliable control limit was calculated based on all the data points, and the new median value was 1.334 mg/kg with a standard deviation of 0.154 mg/kg. The decrease in standard deviation suggested that the precision was improved with smaller data fluctuations. In conclusion, calculation of new and more reliable control limit from more data is useful to find obvious systematic errors, abnormal data and slow changing trend in a timely and effective way and can provide evidence to decide whether data should be retained or removed so as to directly reflect the stability and tendency of the detection.

Sample pretreatment is a key step in element testing, which directly affects the precision and accuracy of analytical results. In order to meet the requirements for the detection of various elements, in this study, we used microwave digestion combined with inductively coupled plasma-mass spectrometry (ICP-MS) operated in the collision reaction mode to determine the content of multiple elements in milk powder, and investigated the influence of sample pretreatment conditions (acid removal time, sample size, and the amount of residual acid) on the detection results. Compared to removing acid mist only, better results were obtained under the following conditions: acid removal temperature was 160 ℃, acid removal time was 97.5 min, and sample size was 0.2500 g, and the amount of residual acid was 1.0 mL. This pretreatment process is characterized by good stability, high precision and high accuracy.