Chemical characterization of particles on wafer

Problem statement: Chemical characterization of particles on a wafer is challenging due to difficulties in (1) locating the defect on the wafer based on KLARF coordinates, (2) acquiring molecular fingerprint spectra from individual particle defects, (3) interpreting the spectra to find the source of particle.


UNISERS advantage: UNISERS defect inspection and classification tool can address these issues thanks to enhanced Raman signals of the particles on the wafer and the intelligent integration of the data analysis software.


Example case: Fabs use ultrafiltration membranes that have sub-10nm pore size. So, they must remove all the particles larger than 10nm. Yet, there are still significant particle defects on the wafer. So, where are these large particles originating from? A reason for large particle (100-200nm) deposition on the wafer is the aggregation of sub-10nm organic impurities (particle precursors). These organic particles are so small that the filters can not remove them. So, after filtration, they agglomerate on the wafer creating larger particles.

On-wafer defectivity analysis of process liquids

Problem statement: The semiconductor production process involves dozens of high-purity process chemicals that can generate particle defects on wafers. Unfortunately, existing liquid analysis techniques can not directly assess the risk of a process chemical regarding creating particle defects on a wafer.


UNISERS advantage: Combining with the successive-spin-drying (SSD) method, we can perform on-wafer particle detection for process chemicals with state-of-the-art sensitivity and accuracy. In addition, the methodology is practical and cost-effective so that on-wafer analysis is feasible in the entire semiconductors ecosystem.


Example case: Three samples of process chemicals are compared regarding their risk of creating particle defects on a wafer. Sample 2 (with red dots) is the best one making the least number of defects on a wafer.

Correlation of liquid defectivity with process defectivity

Problem statement: Single wafer cleaning process involves various process chemicals. A failure in a component (e.g., filter, valve, pipe, container, etc.) of one of the process chemicals can create particle excursion, increasing the random defect count on a wafer. Finding the source of the problem is generally very challenging and time-consuming.


UNISERS advantage: UNISERS can provide both physical and chemical characterization of defects on (1) 300mm test wafers contaminated by particles of the entire process and (2) successive-spin-drying (SSD) wafers contaminated by a specific process chemical. Therefore, utilization of the UNISERS dataset can facilitate a direct correlation of the particles on a wafer with their sources.

Supplementing Surfscan with sub-10nm sensitivity

Problem statement: The detection limit of existing optical defect inspection tools for unpatterned wafers can be hampered due to (1) surface roughness and (2) optical properties of the bottom substrate.


UNISERS advantage: Sub-10nm detection limit of UNISERS is not influenced by the optical properties of the bottom substrate or surface roughness (up to 10A RMS roughness).


Example case: A test wafer (TEOS) for the post-CMP cleaning process is scanned by Surfscan SP5. SP5 could not detect significantly increased defectivity (sub-40nm) on the central part. However, UNISERS scans clearly show increased defectivity for sub-40nm particles at the center of the wafer.

Retention characterization of filters

Problem statement: Ultrafiltration membrane is a critical component in the fabs. Today’s state-of-the-art filters have sub-10nm pore size to filter even the tiniest particle from process chemicals. However, existing metrology tools can not assess the retention efficiency of a filter regarding on-wafer particle defectivity.


UNISERS advantage: Combined with the successive-spin-drying (SSD) method, we can assess the retention efficiencies of filters regarding on-wafer particle defectivities.

Size characterization of CMP slurries

Problem statement: Characterization of the particle size distribution and the large particle count of CMP slurries is critical to tuning CMP and post-CMP cleaning processes.

UNISERS advantage: UNISERS can provide size distribution of particles on the wafer between 8nm and 550nm, providing a practical characterization of CMP slurries

Would you like to test your wafer of process chemical with UNISERS? Please contact