Quantitative native speciation of ppb-level metals in semiconductor-manufacturing-used strong acids and a base

The presence of metal species in solvents significantly impacts production yields in the semiconductor industry, particularly as the dimensions of integrated circuits continue to decrease. Therefore, it is imperative to control metal concentrations in solvents to levels as low as a few parts per billion (ppb) throughout fabrication processes. Effective purification methods are essential for removing various levels of contamination, and understanding the speciation of metals is crucial for achieving efficient purification. Conventional methods for the speciation of solution-phase metals include ion chromatography (IC) and ultraviolet-visible (UV-Vis) absorption spectroscopy. However, these techniques present limitations; for instance, IC can inadvertently alter species during the elution process, while the requirement for high-purity parts per million (ppm) concentrations of metals obscures the speciation of trace mixed samples using UV-Vis absorption spectroscopy. In this study, we present a quantitative speciation method for metals in their native states within strong acids and a base, utilizing the breakthrough curve (BTC) theory in conjunction with inductively coupled plasma-mass spectrometry (ICP-MS). Sodium, potassium, magnesium, calcium, iron, and copper serve as model systems for our investigations. The combination of BTC and ICP-MS provides insights into the species present and their respective abundances. Our findings indicate that breakthrough time (t_BT) is predominantly influenced by the charge states and binding selectivity of the metal species and the concentrations of competing binding species. For scenarios where the product of the adsorption equilibrium constant (K) and the concentrations of a species at equilibrium (C) is significantly less than one (KC ≪ 1), t_BT serves as a critical metric for assessing metal species at trace levels. Taking sodium (I) and potassium (I) at 10 ppb as representative examples, we discovered that t_BT was accelerated by a factor of 5.7 when the concentration of the competing binding species ([H]⁺ in this study) was increased five-fold from 0.02 M to 0.1 M nitric acid (HNO₃). Specifically, the t_BT for sodium (I) decreased from 23 min to 4 min, while for potassium (I), it dropped from 114 min to 20 min. Furthermore, in the cases of magnesium (II) and copper (II) at 10 ppb, t_BT was expedited by a factor of approximately 25; the t_BT for magnesium (II) fell from 100 min to 4 min, and for copper (II), it decreased from 157 min to 6 min when the [H]⁺ concentration was increased five-fold from 0.1 M to 0.5 M HNO₃. Additionally, we observed distinct species transformations for iron and copper, evidenced by markedly altered t_BT in 0.1 M choline hydroxide solutions, which was observed to be less than 10 min. Anionic iron complexes and neutral copper particles were inferred, supported by ion exchange and UV-Vis absorption spectroscopic measurements. Furthermore, copper particles, potentially identified as copper (II) hydroxide or copper (II) oxide, exhibited a size distribution ranging from 200 to 400 nm with a peak at 300 nm, as characterized using particle analyzers. The advantages of the BTC theory-facilitated native quantitative speciation are anticipated to enhance informed decision-making for optimizing purification processes within the semiconductor industry.