Evaluating polymer wear and particulation for semiconductors and data storage - Solid State Technology
This article is from

Evaluating polymer wear and particulation for semiconductors and data storage


(November 22, 2010) -- An overview of wear testing of polymers, with discussion of issues relevant to the semiconductor and data storage industries, is presented by Jeffrey A. Galloway and Sanjiv Bhatt, Entegris Inc. Popular wear test geometries are described and reviewed. Key considerations for designing a wear test, including potential testing issues, are provided. Metrics for evaluating wear are discussed. A test method and accompanying data analysis strategy for evaluating wear of polymers that focus on better serving the semiconductor and data storage industries are recommended.

Although polymers have long been used in products for the semiconductor and data storage industries, such as wafer shippers, reticle pods, process carriers, and read-write head trays, they are susceptible to wear and particulation. Wear and particulation may be caused by contact between a wafer or disc and the polymer product (i.e. a disc sliding into a process carrier) or contact between the polymer product and process equipment (i.e. a process carrier sliding on a conveyor). In addition, alignment and/or insufficient gaps between a material and the polymer product affect the extent of wear and particulation. Since many popular wear test methods were designed for mechanical applications, they do not adequately address the needs of the semiconductor and data storage industries. Most methods focus on wear rate and coefficient of friction (COF), but not on the number and size of particles, a result important to these industries. Since wear particles can contaminate sensitive microelectronics products, reducing yield and product lifetime, it is imperative to understand and minimize the wear of polymers used in these applications. The authors review wear testing of polymers, with emphasis on the needs of the semiconductor and data storage industries. A test method and data analysis strategy for evaluating wear of polymers that focus on better serving these industries are recommended.

Wear test methods

Many factors must be considered when designing a wear test to ensure that the test results are relevant to the final application. The first step in designing a wear test is test geometry selection. Although numerous test geometries have been developed for evaluating the wear performance of polymers, [1] some were designed for mechanical applications, limiting their usefulness for semiconductor and data storage applications. Wear test geometries commonly used for polymers are reviewed below.

Review of wear test geometries

In a pin-on-disc test,[2] a pin with a flat or hemispherical tip contacts a rotating disc. The sample material can be either the pin or the disc. Disadvantages include the difficulty of achieving good alignment for flat pins (necessitating a running-in period to ensure conformal contact) and time-varying contact area for hemispherical pins. The pin-on-disc is our recommended test geometry.

In a pin-on-plate test, [3] a pin with a flat or hemispherical tip contacts a linearly reciprocating plate. The sample material can be the pin or the plate, but the choice can affect wear rate.[4] Disadvantages include the variable sliding velocity and the same pin issues as pin-on-disc.

In a block-on-ring test, [5,6] a block of material contacts a rotating shaft. The sample material is typically the block. Disadvantages include a running-in period to ensure conformal contact between the block and the ring, and the time-varying contact area.

In a disc-on-wheel test,[7] freely rotating abrasive rubber wheels contact a rotating sample disc. Disadvantages include the aggressiveness of the test that often obscures material differences, and accurate measurements require considerable wear. Although data for this popular test is often reported on polymer data sheets, the disadvantages limit its usefulness for semiconductor and data storage applications.

In a thrust washer test,[8] a sample with a washer shape is rotated against a flat plate. Disadvantages include a running-in period to ensure conformal contact between the washer and the plate. This popular test is often used to determine the mechanical limits of the sample material, making it most relevant to mechanical applications involving high pressures and velocities.

Other wear test geometries include ball-on-prism[9] and four ball.[10] These tests are more relevant to mechanical applications such as ball bearings. Figure 1 shows common geometries for wear testing.

Figure 1. Common wear test geometries: (a) pin-on-disc with circular track (b) pin-on-disc with spiral track (c) pin-on-plate (d) block-on-ring (e) disc-on-wheel (f) thrust washer (g) ball-on-prism (h) four ball. Pin-on-disc and pin-on-plate may be identified as ball-on-disc or ball-on-plate, respectively, if the pin has a hemispherical tip.

Other test design considerations

Many wear test geometries have the option of a multi-pass or single-pass configuration. In a multi-pass test, the sample repeatedly passes over the same area of the substrate. An advantage of multi-pass tests is that they can run indefinitely on the same substrate (counterface), allowing an arbitrary amount of wear to be achieved. This advantage has resulted in almost exclusive use of multi-pass configurations for wear testing. However, multi-pass tests have the disadvantage that they are more susceptible to sample heating during testing. In addition, a multi-pass test typically results in material transfer from the sample to the substrate (as debris and/or a transfer film), causing the COF and wear rate to change during testing.

In single-pass tests, the sample continuously contacts fresh substrate. Their main advantage is that the effects of debris and/or a transfer film are eliminated. This leads to a more stable COF and wear rate throughout the test. Samples are less susceptible to heating since the test duration is limited by the substrate area. However, the limited test duration is a disadvantage because samples often have relatively small amounts of wear from sliding against a single substrate. To increase the amount of wear, the test can be extended by changing the substrate. Although wear tests are typically performed with a multi-pass configuration, the choice of configuration depends on the application. Figure 2 shows a comparison of wear and COF trends for single-pass and multi-pass tests.

Figure 2. Comparison of wear and COF trends for (a) single pass and (b) multi-pass tests. Note that the wear rate and COF may increase or decrease in multi-pass tests, depending on the test conditions and material.

Since substrate choice and properties affect the wear mechanism, wear rate, and COF, the substrate should be selected to replicate end-use conditions. The method for preparing the substrate should be kept consistent to ensure repeatable results.

The normal force and sliding velocity should match end-use conditions. However, wear tests often use larger normal forces and velocities to accelerate wear and reduce testing time, especially for materials that have low wear rates under end-use conditions. A disadvantage of accelerated testing is that the material response may differ from the response under end-use conditions. The effect of accelerated testing on the material response may not be universal for all materials evaluated, resulting in negative consequences such as incorrect ranking of materials or meaningless data. For example, owing to the low thermal diffusivity of polymers, they are susceptible to dramatic increases in temperature during accelerated wear testing. If the temperature exceeds the softening point of the polymer, gross melt failure (GMF) can occur. GMF is characterized by a rough and distorted wear surface. Since GMF is a catastrophic failure, its occurrence generally means that the test results are not useful for predicting performance in the end-use application. Figure 3 shows optical micrographs of samples exhibiting normal wear and GMF.

Figure 3. Optical micrographs of wear samples exhibiting (a) normal wear and (b) gross melt failure.

Temperature can influence the tribological behavior of polymers, especially near temperatures at which the polymer softens or melts. Thus, wear tests should be performed at temperatures matching the end-use application. As discussed above, the effect of frictional heating on the sample temperature should be considered.

Other factors that may influence wear test results include relative humidity and crystallinity. The relative humidity should be controlled, especially when testing hygroscopic polymers, to minimize sample-to-sample variability. For semi-crystalline polymers, the crystallinity of the polymer can affect wear test results. Although their effects are typically small, these test parameters should be controlled to match the end-use application.

The method for quantifying wear is also an important factor in designing a wear test. The test duration should be selected such that resolution of the wear measurement method is small relative to the measurement.

Wear data analysis

Common metrics for reporting wear performance are total mass loss, total volume loss, volumetric wear rate, and specific volumetric wear rate. Volumetric wear rate is the volume loss per unit sliding distance (mm3/m). Specific volumetric wear rate is the volumetric wear rate divided by normal force (mm3/[N-m]). Using volume instead of mass for reporting wear performance simplifies comparison of materials with different densities. COF data reported with wear data can be helpful for identifying transient and steady-state wear behavior and for understanding the effects of transfer layers and debris on the wear results.

Optical micrographs of the sample and substrate are often included with wear test results to show transfer layers and debris, and to show the extent of wear for the sample and substrate. Although typically qualitative, analysis of debris is often included with wear test results. Particle size and shape can provide insight into the wear mechanism. For the semiconductor and data storage industries, the number and size of particles is important to determining the effect on product yield and lifetime.

Recommended wear test and data analysis methods

Our recommendations for wear test method and data analysis are presented in Tables 1 and 2, respectively. Methods for calculating wear volume for a hemispherical pin are shown in Table 3.

 

Table 1. Recommended wear test method characteristics.
Characteristic Recommendation Rationale
Test geometry Pin-on-disc 

Constant velocity
Single-pass and multi-pass options
Flexibility in choosing substrate (counterface)

Pin geometry Hemisphere Slight sample misalignment does not affect contact geometry, eliminating need for running-in the sample to establish conformal contact
Since no running-in process is needed, surface tribological properties can be tested
Volume loss (and specific wear rate) can be measured by several methods, offering flexibility in data analysis
Pin material Sample Can determine wear rate without the effect of gross sample deformation (that may occur with a hard pin and soft sample)
Test configuration Single-pass or Multi-pass Selection depends on which configuration is most relevant to final application


Table 2: Recommended wear test data analysis
Metric Rationale
Specific wear rate Easy to compare materials with different densities
Determine if material is low wear
Coefficient of friction Identify events that cause the wear rate to change
Sample inspection by optical microscopyDetermine if GMF occurs
Substrate inspection by optical microscopy Identify transfer layers
Determine size and shape of wear debris
Total number of particles  Determine if material is low particulation
Particle size distributionDetermine how many particles fall in critical size range

Table 3. Methods for calculating wear volume for a hemispherical pin with radius r.
Measurement Calculation
Linear wear or change in sample height, h 
Wear scar diameter, dscar 
Mass loss, Δm

The number and size distribution of particles complements the specific volumetric wear rate, allowing materials with low specific wear rate (low wear materials) to be distinguished from materials that generate a low number of particles (low particulation materials). This distinction is critical in the semiconductor and data storage industries since the number of particles and their size determine the effect on microelectronics. Figure 4 shows wear debris from poly(ether ether ketone) (PEEK), a PEEK-carbon nanotube blend (PEEK-CNT), and a PEEK-carbon fiber blend (PEEK-CF), illustrating the distinction between low wear and low particulation materials. The PEEK and PEEK-CNT particles are much larger than the PEEK-CF particles, but more particles are generated from PEEK-CF. Thus, although the total volume loss for PEEK-CF is less than PEEK or PEEK-CNT, it can generate more particles.

Figure 4. Optical micrographs of wear debris for (a) natural PEEK, (b) PEEK-CNT, and (c) PEEK-CF. The PEEK and PEEK-CNT samples generated fewer, larger particles with larger total volume (low particulation) while the PEEK-CF sample generated many smaller particles with smaller total volume (low wear).

Conclusion

The wear and particulation properties of polymers are critical to their successful application in products used by the semiconductor and data storage industries. Key aspects of designing tribological tests relevant to these industries were reviewed. We identified pin-on-disc as the preferred geometry. Although infrequently used, the single-pass configuration can provide unique information about wear and particulation, and may be the most relevant configuration for certain applications. When evaluating the wear performance of materials for these industries, it is crucial to understand not only the wear rate, but also the COF, particle size, and number of particles generated.

Acknowledgments
We would like to thank the following people for their assistance: Rich Hoffman (blend preparation), Mingjun Yuan (injection molding), the Sample Machining Department (sample preparation), and Yingkai Liu/Technology Characterization Laboratory (sample testing).

References
1.  ASTM Standard G190-06, ASTM International, West Conshohocken, PA, 2006.
2. ASTM Standard G99-05, ASTM International, West Conshohocken, PA, 2005.
3. ASTM Standard G133-05, ASTM International, West Conshohocken, PA, 2005.
4. S. M. Hosseini and T. A. Stolarski, Surf. Eng. 4 (1988) 322-326.
5. ASTM Standard G137-97(2009), ASTM International, West Conshohocken, PA, 2009.
6. ASTM Standard G176-03(2009), ASTM International, West Conshohocken, PA, 2009.
7. ASTM Standard G195-08, ASTM International, West Conshohocken, PA, 2008.
8. ASTM Standard D3702-94(2009), ASTM International, West Conshohocken, PA, 2009.
9. O. Jacobs, N. Mentz, A. Poeppel, K. Schulte, Wear 244 (2000) 20–28.
10. O. Jacobs, R. Jaskulka, C. Yan, W. Wu, Trib. Lett. 18 (2005) 359-372. 

Jeffrey A. Galloway received his B.S. in Chemical Engineering at the University of Wisconsin-Madison and his Ph.D. in Chemical Engineering at the University of Minnesota-Twin Cities. He is a Research Scientist at Entegris, Inc., 101 Peavey Road, Chaska, MN 55318; 952-556-4058; [email protected]  

Sanjiv Bhatt received his Ph.D. in Polymer Engineering and is a Technical Director and Principal Scientist at Entegris, Inc.

Subscribe to Solid State Technology/Advanced Packaging. Follow Solid State Technology on Twitter.com via editors Pete Singer, twitter.com/PetesTweetsPW and Debra Vogler, twitter.com/dvogler_PV_semi. Or join our Facebook group

Advertisement