STI CMP post cleaning removes abrasive residues and organic layers to enable subsequent deposition uniformity .
The Shallow Trench Isolation (STI) module provides critical electrical isolation between active device regions, preventing parasitic crosstalk and leakage currents in advanced CMOS arch
itectures . Immediately following the STI Chemical Mechanical Planarization (CMP) step, the wafer surface is covered with abrasive residues, pad debris, and organic passivation layers . This STI CMP Post Cleaning step is specifically designed to completely remove these complex residues before the subsequent STI Final Densification Anneal and Wet Deglaze Etch . Unlike backend-of-line (BEOL) post-CMP cleans that must carefully balance metal corrosion against particle removal, this front-end step focuses exclusively on clearing dielectric surfaces (SiO2 and Si3N4) of highly reactive ceria (CeO2) particles and robust organic inhibitors . Failing to achieve a pristine surface here will cause residual organics to carbonize during the subsequent high-temperature anneal, while remnant particles will act as micromasking defects during the subsequent SiN strip step . During the preceding STI CMP step, ceria-based high-selectivity slurries are widely utilized due to their unique chemical affinity for SiO2, which involves the formation of reversible Ce–O–Si bonds . While this strong chemical reactivity enables efficient oxide planarization and high selectivity against the underlying silicon nitride stop layer , it also causes ceria particles to bind aggressively to the wafer surface . Simultaneously, the high selectivity is often achieved by employing organic additives, such as nitrogen-containing aromatic heterocyclic compounds or anionic surfactants like SDS, which selectively adsorb onto the Si3N4 surface to form a dense passivation layer [A1, P3]. The core cleaning mechanism relies on the synergistic application of strongly oxidizing chemical solutions, such as sulfuric acid-hydrogen peroxide mixtures (SPM), which effectively dissolve the metallic oxide particulates and oxidatively degrade the tightly bound organic inhibitor films . This chemical dissolution is typically coupled with controlled fluid dynamics and physical detachment mechanisms, overcoming the van der Waals and electrostatic adhesion forces between the nanoscale contaminants and the dielectric substrates . The selection of harsh oxidizing chemistries like SPM is necessitated by the chemical inertness of the cerium oxide residues and the high thermal stability of the boundary-lubricating surfactant films . To process substrates efficiently without introducing temperature fluctuations or bath concentration drops, advanced cleaning systems utilize closed-environment single-wafer or quasi-continuous processing tanks with controlled isolation gaps between substrates . The efficiency of particle removal and organic stripping is highly dependent on the interactions between chemical concentration, fluid temperature, and substrate residence time . Increasing the bath temperature exponentially accelerates the oxidative degradation of organic additives but concurrently accelerates hydrogen peroxide decomposition, requiring precise replenishment controls to maintain the oxidation-reduction potential (Engineering Practice). Furthermore, adjusting the pH during subsequent rinse steps is critical to modulating the zeta potential of both the particles and the dielectric surfaces, ensuring mutual electrostatic repulsion that prevents particle re-adsorption . In the context of a nanoscale Backside Illuminated (BSI) CMOS Image Sensor flow, the tolerance for post-CMP defects is exceptionally stringent . Residual ceria particles or micro-scratches generated during the planarization process directly compromise the integrity of subsequent gate oxides . Because ideal Si-SiO2 interfaces are the main factor responsible for the success of MOSFET performance , any localized interfacial disruption caused by uncleaned residues can induce severe electrical variability . In image sensors specifically, such defects in the active area or isolation trenches act as mid-gap generation-recombination centers, resulting in elevated dark current and "hot pixel" failures that degrade overall image quality (Engineering Practice).
Sign in to continue through all 417 steps