在小体积（< 50 μL）液体取样时, 容易产生较大的取样误差, 为了降低分析结果的不确定度, 常常用质量定量法代替体积定量法。 传统标准加入法以样品体积定量, 不能用于以样品质量定量的场合。 为此, 我们提出了一种变形标准加入法, 以便用于以样品质量定量的场合。 以ICP-OES法测定复杂溶液体系中低含量及微量元素Hg, Mo和Rh为例, 对变形标准加入法进行了介绍。 标准加入法的目的是为了校正溶液基体效应, 而溶液基体效应包括两种不同类型的干扰: “恒定干扰”和“比例干扰”。 变形标准加入法只能校正溶液基体效应中的“比例干扰”, “比例干扰”的大小可以用一个定量指标k表示: 当k=1时, 表示不存在“比例干扰”的影响; k偏离1越远, 则“比例干扰”的影响越大。 至于“恒定干扰”的影响, 则可以利用仪器自身的背景校正方法予以降低或消除。 变形标准加入法测定结果不确定度主要来自于背景校正, 与所选分析线的信背比密切相关。 信背比越低, 背景校正的不确定度越高, 因此实际分析中应尽可能选择具有较高信背比的分析线, 否则, 即使事先经过了背景校正, 最终的分析结果也可能包含很大的误差。

When a small volume (<50 μL) of liquid is sampling, it is easy to produce large sampling error. In order to reduce the uncertainty of analytical results, the quantification based on the sample volume is usually replaced by the quantification based on the sample mass. The conventional standard addition method (SAM), in which the quantification is based on the sample volume, cannot be applied when the quantification is based on the sample mass. Therefore, a modified SAM with mass-based quantification is proposed. As a demonstration, the modified SAM has been applied to determine low or trace elements Hg, Mo and Rh in complex solution with ICP-OES. The aim of the SAM is to correct the matrix effect, while the matrix effect includes two kinds of interferences: constant interference and proportional interference. Only the proportional interference can be corrected with the modified SAM, which can characterize the extent of the proportional interference with a quantitative index k: k=1 means there is no proportional interference; while the larger difference between k and 1 means greater proportional interference existing in the sample. As for constant interference, it can be reduced or eliminated by applying the background correction function of the ICP-OES instrument. The uncertainty of analytical results mainly comes from background correction, and depends on the signal-to-background ratio of the analytical line selected. The lower the signal-to-background ratio is, the greater the uncertainty of background correction is. Therefore, in the actual analysis, we should try to choose the analytical line which has as high signal-to-background ratio as possible. Otherwise, the final analytical result maybe contains very large uncertainty, even though the background correction has been made in advance.