考虑到市售氢化物发生-原子荧光光谱仪载气由石英原子化器内管载入，载气流量较小（300~600 mL·min^-1，测定铅和镉），产生的铅、镉荧光强度较小，因此改造了此仪器的石英原子化器中载气和辅助气的流路，将内管进样改为外管进样，同时优化了助剂及载流，并进行了方法学验证。样品经酸溶液处理后，用改造的氢化物发生-原子荧光光谱仪分析。结果显示：相较于内管进样，外管进样可将载气流量提升至1 000~1 400 mL·min^-1，载带的氢化气更多，氢化气和氢气混合气能更快地与石英套管上端的加热炉丝接触而点燃；内管输入辅助气（氩气），会将氢火焰焰炬向上适度抬升，形成的焰炬更稳定，体积也更大。以铁氰化钾和镉增敏剂作助剂，可以显著增强铅、镉荧光强度；以水替代传统稀酸作载流，交替进样时试验空白中铅、镉荧光强度的相对标准偏差（RSD，n=7）均小于3.0%，适用于小批量样品检测；铅、镉标准曲线的线性范围分别在10.00 μg·L^-1和0.50 μg·L^-1以内，检出限（3s/k）分别为3.5 μg·kg^-1（膨化食品中铅）和0.74 μg·kg^-1（大米中镉）。铅、镉灵敏度（标准曲线斜率k）分别是文献报道同类方法的1.3~13倍和2.8~99倍。按照标准加入法进行回收试验，回收率为86.0%~118%，测定值的RSD （n=6）为2.6%~5.8%。

Considering that the carrier gas of the commercially available hydride generation atomic fluorescence spectrometer was loaded into the inner tube of the quartz atomizer with a small flow (300-600 mL·min^-1 for the determination of lead and cadmium), resulting in smaller lead and cadmium fluorescence intensities, the flow paths of the carrier gas and auxiliary gas in the quartz atomizer of this instrument had been modified, and the inner tube injection had been changed to the outer tube injection. Meanwhile, the additives and carrier current had been optimized, and methodological validation had been conducted. After being treated with acid solution, the sample was analyzed using the modified hydride generation atomic fluorescence spectrometer. It was shown that compared to the inner tube injection, the outer tube injection could increase the carrier gas flow to 1 000-1 400 mL·min^-1, carrying more hydrogenated gas. The mixture of hydrogenated gas and hydrogen gas could come into contact with the heating furnace wire at the upper end of the quartz sleeve faster and beignited. The auxiliary gas (argon) input from the inner tube would moderately lift the hydrogen flame upwards, resulting in more stable and larger volume of the flame. Using potassium ferricyanide and cadmium sensitizer as additives, the fluorescence intensities of lead and cadmium could significantly be enhanced. Replacing traditional dilute acids with water as carrier current, RSDs (n=7) of fluorescence intensities of lead and cadmium in the reagent blank during alternating injection were less than 3.0%, suitable for small batch sample detection. Linear ranges of lead and cadmium standard curves were within 10.00 μg·L^-1 and 0.50 μg·L^-1, with detection limits (3s/k) of 3.5 μg·kg^-1 (lead in puffed food) and 0.74 μg·kg^-1 (cadmium in rice). The sensitivity (standard curve slope k) of lead, cadmium was 1.3-13 and 2.8-99 times those of similar methods reported in literatures, respectively. Test for recovery was made according to the standard addition method, giving recoveries in the range of 86.0%-118%, and RSDs (n=6) of the determined values were found in the range of 2.6%-5.8%.