IR-PiFM: Molecular Detection and Identification of Nanoscale Surface Contamination
Introduction: With continuous advances in semiconductor fabrication in branches such as high-NA/hyper-NA lithography and advanced packaging, controlling and characterizing nanoscale residual contamination after cleaning has become a critical yield and reliability challenge. These trace residues can be difficult to identify with conventional surface analysis techniques due to their nanoscale distribution and chemical complexity. This work demonstrates the application of Infrared Photo-Induced Force Microscopy (IR-PiFM) as a high-resolution, chemically specific technique to detect and identify adhesive-related contamination on a post-clean wafer surface used for hybrid bonding. PiFM combines atomic force microscopy with infrared spectroscopy, enabling simultaneous nanoscale imaging and vibrational chemical identification with sub-5 nanometer spatial resolution and monolayer sensitivity.
Experimental: Measurements were carried out on two 300 mm through-silicon via (TSV) wafers, one fresh wafer pre-temporary bonding (TB), and one after debonding (DB) and cleaning; a silicon wafer coated in a thin film of adhesive was also analyzed as a reference. All PiFM measurements were carried out using a Vista 300 microscope coupled with an optical parametric oscillator with a difference frequency generator as the laser source.
Results: IR-PiFM spectra on the post-DB and cleaning sample show the same characteristic peaks found on the adhesive reference at 2915, 2850, 1460, 1375, and 700 cm-1 (Fig. 1). The first four peaks are also found in various C-H contaminants; therefore, 700 cm-1 is used as a unique peak for adhesive detection for these measurements. IR-PiFM imaging enables differentiation between adhesive residues, background hydrocarbons, and underlying Si substrate or Cu via; the adhesive matches the small bumps on the surface topography, and it is found across the Cu vias and Si substrate (Fig. 2). The adhesive is less than 10 nm thick, and, in most cases, these bumps are only ~3 nm thick. IR-PiFM also differentiates these from the slightly larger bumps on Cu, which are Cu2O growths and found on both TSV samples, but most notably the pre-TB spectrum (Fig. 1, red), and are indicated by the absorption peak around 620 cm-1.
Conclusion: The results highlight PiFM’s value as a diagnostic and failure-analysis tool for surface preparation and cleaning optimization in advanced packaging. By enabling nanoscale molecular identification and mapping of residual contamination, PiFM supports improved process control and more robust qualification of cleaning strategies for next-generation semiconductor manufacturing.
BIOGRAPHY
Padraic O’Reilly joined Molecular Vista Inc. (MVI) in 2017, initially as an Applications Engineer before moving into his current role as an Applications Scientist, where he works closely with customers to develop experiments utilizing the AFM techniques available with the MVI’s instruments, including Electrostatic Force Microscopy (EFM), Kelvin Probe Force Microscopy (KPFM) and, primarily, Photo-induced Force Microscopy (PiFM). During his time with Molecular Vista, Padraic has garnered vast knowledge in the visible and IR PiFM techniques through thousands of hours of hands-on experience with an extensive array of specimens, including contaminant and defect review for the semiconductor industry.