Detecting Large Particles in Chemical Mechanical Polishing (CMP)

Polishing with the help of CMP slurries is an integral part of microcircuit manufacturing. The particle size distribution (PSD) measurement on the CMP slurries is required to ensure their health, which is a critical parameter to improve the device yields.

Besides measuring the mean size of the PSD, the monitoring method must be able to detect the presence of tails caused by contamination or aggregation of particle because of CMP delivery systems or changes in chemistry, or by applied shear forces. The ideal characterization system must be able to provide an accurate large particle count (LPC) value.

Particle Size/Counting Techniques

The particle size and distribution data can be obtained using the light scattering techniques such as laser diffraction and dynamic light scattering methods, but not any meaningful concentration data.

Particles are measured individually by single particle optical sizing (SPOS) by passing them through a narrow measurement chamber, thus providing accurate size and concentration data. The high resolution of SPOS enables detecting individual particles that are stripped from the main distribution, making it suitable to detect the LPCs affecting the health of the CMP slurries. The SPOS is both a laboratory and a point-of-use (POU) instrument.

The AccuSizer Range of Systems

Large particle tails are efficiently determined with the help of the Particle Sizing Systems AccuSizer, which can perform measurements at full concentration or with automatic dilution to improve experiment conditions. Figure 1 shows the new PSS AccuSizer FX system (POU system) can measure sizes of smaller particles at higher concentrations.

The new PSS AccuSizer FX system

Figure 1. The new PSS AccuSizer FX system

The new FX sensor has better concentration limit due to the reduction in the total volume analyzed with the help of a focused beam. As a result, measurements can be made without dilution. The FX sensor can measure a size range of 0.65-20µm at concentrations 200 folds higher than standard light extinction or scattering sensors. It is possible to display the results in up to 512 size channels.

Besides being used as a laboratory or POU instrument, the system can be installed in the FX Tower system to multiplex four CMP slurry streams through the sensor/counter. With all these configurations, the accurate and reliable LPC data can be obtained for CMP slurries. The AccuSizer is a proven system for the determination of LPC > 0.5µm in CMP slurry. In Figure 2, different techniques were compared to determine a known concentration of 1µm SiO2 particles incorporated to a base of silica oxide CMP slurry. The perturbation detection limit of the AccuSizer (older model 780) was found to be at 0.07mg/L.

Perturbation detection analysis: a method for comparing instruments that can measure the presence of large particles in CMP slurry

Figure 2. Perturbation detection analysis: a method for comparing instruments that can measure the presence of large particles in CMP slurry

LPC Detection

In this experiment, 1µm polystyrene latex (PSL) spheres were used to spike a number of common silica CMP slurries to corroborate the ability to determine 1µm particles in CMP slurries. After corroborating for size and concentration, the spike particles were then diluted for the subsequent analyses using the AccuSizer FX POU system featuring an FX sensor with a range of 0.65-20µm.

Result 1

The dilution of Slurry A to 250:1 was done in order to ensure that the measured values were fallen outside the coincidence region. Then the 250mL of sample suspension was added to 1.44mL of 1000 x diluted PSL stock (1.74 × 107/mL) and spiked to a concentration of 100,000/mL. The results are presented in Figure 3, and Tables 1 and 2.

Slurry A before (blue) and after 1µm spike (red)

Figure 3. Slurry A before (blue) and after 1µm spike (red)

Table 1. Slurry A result before spike

Diameter Range (microns) # Part. Sized (#/ml) Cum Num >=Diam. (#/ml)
0.70 - 0.72 1107410 2637543
0.72 - 0.77 738493 1530133
0.77 - 0.82 372885 791640
0.82 - 0.87 203005 418755
0.87 - 0.92 100429       215749  
0.92 - 0.98 54625 115320
0.98 - 1.04 32665 60695
1.04 - 1.11 13392 28030
1.11 - 1.18 7375 14639
1.18 - 1.25 4147 7264
1.25 - 1.33 1762 3117
1.33 - 1.41 709 1355
1.41 - 1.50 213 645
1.50 - 1.60 60 432
1.60 - 1.70 88 372

Table 2. Slurry A result after spike

Diameter Range (microns) # Part. Sized (#/ml) Cum Num >=Diam. (#/ml)
0.70 - 0.72 1005006 2484223
0.72 - 0.77 633343 1479216
0.77 - 0.82 343434 845874
0.82 - 0.87 183631 502439
0.87 - 0.92 96589       318808  
0.92 - 0.98 53409 222219
0.98 - 1.04 82906 168810
1.04 - 1.11 64733 85904
1.11 - 1.18 11916 21172
1.18 - 1.25 4192 9256
1.25 - 1.33 2477 5064
1.33 - 1.41 939 2587
1.41 - 1.50 371 1647
1.50 - 1.60 171 1776
1.60 - 1.70 0 1105

Result 2

Here, the FX POU system was used to measure the Silica slurry B without dilution (Figure 4). The concentration measured was ~ 20,000 particles/mL. Then, 57µL of the 1000 x diluted 1µm PSL standard (concentration = 1.74 x 107 particles/mL) was added to 200mL of the original slurry to spike this sample. Figure 5 shows the results prior to and after the spike.

Slurry B, no dilution, no spike

Figure 4. Slurry B, no dilution, no spike

Slurry B before (blue) and after spike (red)

Figure 5. Slurry B before (blue) and after spike (red)

Result 3

The slurry used in the fab was Slurry C, a dilution of 1 part slurry to 2 parts DI water. The FX POU system was used to measure this base slurry without dilution. Figure 6 shows the results prior to spiking. There is no sharp drop in the LPC tail as observed in Slurry A and B. The concentration was ~ 30,000 particles/ mL > 0.7µm, with a particle count of ~ 5,500 particles counted in 30mL of slurry.

Slurry C, no dilution, no spike

Figure 6. Slurry C, no dilution, no spike

Then, 35µL of the 1000 x diluted 1µm PSL standard (concentration = 1.74 x 107 particles/mL) was added to 250mL of the base slurry to spike the Slurry C. The results for slurry C prior to and after spiking are presented in Figure 7. The spike concentration estimated was to be 2,500 particles/mL.

Slurry C before (blue) and after (red) spike

Figure 7. Slurry C before (blue) and after (red) spike

Conclusion

The results demonstrate the ability of the PSS AccuSizer FX to determine low concentration LPC particles in CMP slurries.

This information has been sourced, reviewed and adapted from materials provided by Particle Sizing Systems, An Entegris Company.

For more information on this source, please visit Particle Sizing Systems, An Entegris Company.

Citations

Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Entegris Particle Sizing. (2019, August 22). Detecting Large Particles in Chemical Mechanical Polishing (CMP). AZoM. Retrieved on April 05, 2020 from https://www.azom.com/article.aspx?ArticleID=11521.

  • MLA

    Entegris Particle Sizing. "Detecting Large Particles in Chemical Mechanical Polishing (CMP)". AZoM. 05 April 2020. <https://www.azom.com/article.aspx?ArticleID=11521>.

  • Chicago

    Entegris Particle Sizing. "Detecting Large Particles in Chemical Mechanical Polishing (CMP)". AZoM. https://www.azom.com/article.aspx?ArticleID=11521. (accessed April 05, 2020).

  • Harvard

    Entegris Particle Sizing. 2019. Detecting Large Particles in Chemical Mechanical Polishing (CMP). AZoM, viewed 05 April 2020, https://www.azom.com/article.aspx?ArticleID=11521.

Ask A Question

Do you have a question you'd like to ask regarding this article?

Leave your feedback
Submit