Coping with Ultrapure Water Particles via Testing Standard with SEMI Document #5188

Coping with Ultrapure Water Particles via Testing Standard with SEMI Document #5188

By Slava Libman (Balazs NanoAnalysis), Marty Burkhart (HiPureTech), David Blackford (Fluid Measurement Technologies), Michael Tran (SEMI Standards)

In April 2011, the SEMI Standards Liquid Chemicals Committee approved the formation of a new task force for Ultrapure Water Filtration Efficiency Task Force looking for a work-around-solution for one of the most critical issues related to process fluid technology – on-line particle (<50nm) control in ultrapure water (UPW) streams. The International Technology Roadmap for Semiconductors (ITRS) defined the risk related to the particle control several years ago when it became clear that the UPW on-line particle metrology was not able to keep pace with semiconductor industry needs. The issue has turned even more critical as UPW filter pore size becomes marginal with respect to the critical particle size and as semiconductor technology becomes even more susceptible to defects.  In short, troublesome particle sizes and allowable quantities exceed the detection limits of today’s on-line particle measuring equipment.

One creative option offered by the UPW ITRS has been a wafer risk mitigation strategy based on off-line validation of the filter performance. This approach uses the ability of existing and alternative metrology to detect particles at high concentrations (much larger than that typically found in UPW). If effective removal of a large number of particles can be demonstrated, it is assumed that such filters would provide effective particle control under normal UPW system conditions.

Through the SEMI Standards Program, a taskforce comprised of a wide range of leading technology companies, including analytical laboratory, particle metrology, filtration, consulting, as well as end users, was formed. The group included representatives of Balazs NanoAnalysis, HiPureTech, Fluid Measurement Technologies, CT Associates, Entegris, GE Water, W.L. Gore, Pall, Siltronic, Lighthouse, and PMS.

Bringing forth a workable standard was a shared goal by all since in addition to the real wafer production risk, there is (1) a need to generate comparison filter performance, thus data allowing selection of the best performing filters, (2) demonstration of the performance of different types of filters, (3) development of a uniform set of testing instructions and (4) finally be able to quantify results instead of relying on filter material characteristics as a qualitative indicator.

After a yearlong effort involving extensive experimental and theoretical work, Document #5188 was submitted for SEMI Standards balloting for Cycle 4 in 2012. The document is a guide as it deals with testing in the area where the performance data is limited and more data is required to fully validate the assumptions and methods used. The key findings that determined the content of the document submitted for balloting are as follows:

Type of the challenge particle is important as different particles are retained by filters with different efficiency. Silica particles are the worst case scenario type.  See Figure 1 where polystyrene latex (PSL) beads show 100% retention by the filter while silica is much less.

Retention of PSL Followed by Silica

Figure 1
Retention of PSL Followed by Silica

During the extensive study, it was learned that concentration of the challenge particles plays a negative role in retention of the challenge particles by filters (see Figure 2). Driven by the metrology limitations, the results of the tests using high challenge particle concentration need to take into account this phenomenon.

Examples Showing the Effect of Particle Concentration on Particle Retention at Fixed Loading

Figure 2
Examples Showing the Effect of Particle Concentration on Particle Retention at Fixed Loading

Challenge test duration is important, as overloading the filter with particles affects the filter performance. The test itself limits the loading to less than one monolayer. As seen in Figure 3, extending the duration of the test to allow 10 monolayers to develop can produce unrepresentative results.

Retention of 28nm Silica Particles by 4 Filter Types

Figure 3
Retention of 28 nm Silica Particles by 4 Filter Types

One of the obstacles the taskforce had to overcome was finding a challenge particle source with relatively uniform particles size distribution. After some research, sources were identified (colloidal silica) within a completely unrelated industry. Following evaluation of a particle size distribution measurement those sources were found to be suitable.  See Figure 4 for the size distribution of a relatively monodispersion of commercially available colloidal silica.

Differential Number Concentration vs. Particle Diameter

Figure 4
Differential Number Concentration vs. Particle Diameter (nm) for Commercially Available Colloidal Silica

The proposed guide had no choice but to use a non-standard challenge material as 10-20 nm particles are not available from NIST or other standard organizations. Finally the document deals with the limitations of the existing particle metrology and why it was necessary to use a new non-standard instrument for the testing.

In summary, particle mitigation is a crucial parameter to monitor to ensure the production of high yielding wafers.  The industry currently has no means of testing filter effectiveness to retain sub 50 nm particles within UPW streams, making Document #5188 both an important and urgently needed guide for end-users worldwide.  The task force recognizes that this is a first attempt effort and plans to make numerous improvement revisions as more knowledge is gained through end-user usage and feedback.

Ongoing Parallel Activity

Under the Japan Liquid Chemicals Committee, the Japan Liquid Filter TF was chartered in 2010 to develop a standard of mono-dispersed polystyrene latex (PSL) challenge test for liquid filter. In July 2011 the TF began its development work on a Test Method for Particle Removal Performance of Liquid Filter Rated 20-50 nm with Liquid-borne Particle Counter.  This activity aims to filter for liquid chemicals rated 20-50 nm that have been widely used, but the standard test method was not addressed well. Therefore, this new standard test method should be established.

Participation in SEMI Standards

If you have any questions regarding SEMI Standards participation or would like to join the Liquid Chemicals committee or the Ultrapure Water Filtration Efficiency Task Force please contact SEMI at

The SEMI Standards Membership Application Form is available at:

SEMI Standards Watch, July 2012