Metal contamination in photoresists, developers and lithography solvents is a longstanding concern of the semiconductor industry since both dissolved metals and metal-containing nanoparticles can produce defects during the many repeated lithography steps that are used to create integrated circuits. These lithographic materials directly contact the wafer during processing, causing each step to become an opportunity to transfer trace metallic impurities to the wafer surface.1,2
Detecting unintended metals in a lithography material requires understanding whether metal contamination was already present in the material or was introduced somewhere along the sampling and analytical workflow. Removal also requires identifying a contaminant’s form because the purification method needs to be tailored to both the metal species and the contaminated material.
Where contamination originates
One of the most commonly overlooked sources of metal contamination is not the sample itself, but rather the items that come into direct contact with it during sample preparation and handling. Pipette tips, sample containers, caps, tubing, transfer tools, and even gloves can all become meaningful contamination sources when the target concentration is at the ppt or sub-ppt level.
“These items are often assumed to be clean because they are disposable, made from plastic, or used routinely in the laboratory. However, in ultra-trace semiconductor analysis, even a very small amount of leaching from a container, residue from a pipette tip, contamination introduced during bottle opening, or particles generated during handling can affect the final result,” said Kazuhiro Sakai, ICP-MS Product Manager in Agilent’s Spectroscopy Solutions Division.
Pipette tips are a good example. They are in direct contact with the sample or standard solution, but are often treated as a low-risk item. In practice, each new tip can introduce a new contamination risk, especially if it has not been properly selected, rinsed or controlled for ultra-trace analysis.
“The best way to reduce this risk is to minimize the number of sample handling steps,” said Sakai. “Users should select low-contamination materials such as PFA where appropriate, avoid ordinary glassware for ultra-trace metal analysis, and clean or rinse labware and consumables with suitable high-purity reagents before use. Preparing samples and standards by weight rather than by volume can also reduce reliance on volumetric glassware and repeated pipetting. Where possible, automation and direct sampling from the original container can further reduce operator contact and lower the risk of contamination.”
Ultra-trace metal detection techniques
As contamination tolerances tighten, metallic impurities in lithography materials need to be measured at ppt or sub-ppt levels. ICP-MS is the main technique for identification and quantification of metallic contamination in lithography materials at these levels.
“ICP-MS and ICP-MS/MS are powerful tools for measuring trace and ultra-trace metal concentrations, especially when spectral interferences and background signals must be controlled,” said Sakai.
This technique is especially powerful for complex photoresist and solvent samples because it can reduce spectral interferences that would compromise ultratrace measurements. For example, an Agilent 8900 ICP-QQQ can quantify 20 trace elements in semiconductor photoresists diluted tenfold in PGMEA, achieving single- or sub-ppt background equivalent concentrations for all measured elements.3
ICP-MS results may require complementary material characterization techniques, such as graphite furnace atomic absorption spectroscopy (GFAAS) for targeted trace metals in semiconductor-related organic solvents. It is especially useful when only one or a few elements need to be rapidly quantified. In another example, a PerkinElmer PinAAcle 900Z atomic absorption spectrometer was used to determine Na, Ca, Cu, Fe, and Al in photoresist stripper solutions, TMAH, PGMEA, IPA, and NMP, with acceptable spike recoveries across the tested matrices.4 A 2021 study determined 20 trace and ultra-trace elements in semiconductor-grade organic process chemicals.5
Identifying the metal contaminant’s form
It’s critical to establish the form and origin of a detected metal, especially for materials where metals are intentional components, such as metal-oxide photoresists.
“The goal is not simply to ask whether a metal is present, but whether its level, form, and behavior are consistent with the intended material design,” said Sakai. “Unintended contamination is typically identified through unexpected elements, abnormal concentration changes, poor reproducibility, unexplained lot-to-lot variation, or elevated signals in blanks, containers, reagents, or preparation steps.”
It is also necessary to distinguish dissolved metals from particulate contamination. Single-particle ICP-MS (spICP-MS) can be used to confirm the presence of metal-containing nanoparticles, as well as their number, concentration, and size. Agilent uses spICP-MS to determine Fe₃O₄ nanoparticles spiked at 5 ppt in IPA, PGMEA, and butyl acetate, and multielement nanoparticles in TMAH.6
Identifying the metal contamination source
Once a metal is detected, a major challenge is determining whether it originated in the material or was introduced during sampling and preparation.
“In ultra-trace semiconductor analysis, contamination control must include not only the instrument, but also the entire sample path from the original container to the plasma,” said Sakai. Controlling the analytical environment is just as important as minimizing direct handling.
Although cleanroom classification is also important, the local sample exposure within the cleanroom often matters more than the cleanroom classification itself.
“An ideal placement would separate the ICP-MS from sample preparation and chemical handling as much as practical, while keeping the sample path short and well controlled,” Sakai says.
Low-traffic locations, minimized bottle opening and transfers, clean enclosures around autosamplers and contamination-qualified labware can help minimize metal contamination during analysis.
Removing metal contaminants
The removal method has to match the contaminant form and the chemistry of the material without disrupting the working material’s composition.7 In new research Entegris showed that two purifier media remove about 90% and 94% of total added metals from PGMEA, but performance varied substantially by solvent and element. This demonstrates that purifier chemistry must be matched to the specific application, further illustrating the importance of accurately identifying the form of metallic impurities.
Formulated photoresists are more difficult to purify than individual components. Entegris notes that modified membranes and ion-exchange media can remove metals from lithography chemicals, but functional groups on a membrane may also interact with resist components and reduce either metal-removal efficiency or resist performance.8 One strategy is to purify materials before formulation and then use photoresist-compatible filtration for particulate contamination and confirm that the final filtration step does not disturb resist chemistry.
Final thoughts
Accurately identifying metal contamination in photoresists and lithography solvents requires reliable characterization using techniques like ICP-MS to understand the contaminant’s form, source, and potential impact. Once a contaminant is characterized, a carefully selected purification method can address both the final formulation itself and potential sources throughout its handling.
Sakai advises all laboratories to be vigilant.
“The most unexpected source is often the most ordinary part of the workflow,” he concludes.