---

Meet the Presenter
This article is based on a webinar titled Friction and nanoscale wear of stiff multi-contact interfaces. Hosted by the American Society of Mechanical Engineers’ (ASME) Tribology Division and presented by Barb Weber on Nov. 13, 2024, the session explored how surface damage from friction and wear is known to cause significant performance issues and economic losses across various industries. Courtesy of STLE, this article captures the core insights from this ASME-organized event. For more technical content from the ASME Tribology Division Webinar Series, visit the ASME Tribology Division’s website at www.asme.org/get-involved/groups-sections-and-technical-divisions/technical-divisions/technical-divisions-community-pages/tribology-division.

Bart Weber leads the Contact Dynamics group at the Advanced Research Center for Nanolithography and was appointed associate professor at the University of Amsterdam, both in the Netherlands. His research focuses on various aspects of tribology: the science of friction, lubrication and wear. Topics of interest have ranged from understanding how small amounts of water influence friction on granular materials, to interrogating the slipperiness of ice and visualizing the interplay between surface topography, contact mechanics and friction using stress-sensitive fluorescent probe molecules. The emphasis of current work, supported by ERC-StG project CHIPFRICTION, is on understanding how wear and adhesion influence the short stroke friction behavior of stiff materials in the context of nanolithography. You can reach him at b.weber@arcnl.nl.
 

Bart Weber

---

KEY CONCEPTS
• Friction and wear at multi-contact interfaces can strongly influence wafer handling and alignment in semiconductor manufacturing, with repeated versus non-repeated sliding producing disparate outcomes.
• Atomistic processes—including covalent bond formation and single-layer wear events—play a surprisingly important role in determining friction and wear behavior, even in large-scale contacts.
• Understanding how surface chemistry, debris buildup and roughness affect friction provides actionable insights that apply fundamental tribology research to the semiconductor industry. 
 
Friction and wear are critical challenges in semiconductor manufacturing. Components such as wafer stages, lithography clamps and polishing tools must move with sub-micron accuracy. Even minor friction can cause positioning errors, vibration or contamination, leading to defects in chips worth millions of dollars.

Wear compounds the problem by degrading surfaces over time, introducing particle debris into ultra-clean environments where even a single contaminant can compromise an entire batch of wafers. For a high-volume industry where yield is everything, controlling these effects is essential for quality and profitability.

Beyond contamination, friction and wear directly influence tool longevity and manufacturing efficiency. For example, if wafer handling mechanisms or polishing pads degrade too quickly, production faces costly downtime and increased maintenance. Advanced coatings, lubricants and tribological research help mitigate these challenges, allowing tools to run longer, cleaner and more reliably.

This article is based on a webinar by Bart Weber, Advanced Research Center for Nanolithography and University of Amsterdam, and the Tribology Division of the American Society of Mechanical Engineers (ASME). See Meet the Presenter for more information.

ARCNL’s role in bridging fundamental science and industry needs
The Advanced Research Center for Nanolithography (ARCNL) combines academic expertise with industrial collaboration to study friction, wear and contact mechanics in semiconductor systems. Located in Amsterdam’s Science Park, it is a collaborative institute with public and private partners.¹ 

ARCNL is focused on fundamental physics and chemistry that is relevant to future and current key technologies in nanolithography and is thus important to the semiconductor industry. With this foundation, the ARNCL team has designed experiments to uncover how different types of sliding affect wear.

The research centers on friction and wear of stiff, wear-resistant interfaces with an emphasis on how these topics relate to one another. The motivation comes from the semiconductor industry, where companies manufacture chips by packing increasing numbers of transistors into smaller areas. This constant push for miniaturization applies not only to data chips but also to memory chips. As features shrink to just a few nanometers, precise positioning of the substrates used in chip printing becomes increasingly critical. 

In nanolithography machines, the interaction between the backside of a wafer and the surface of the substrate holder involves friction, wear and occasional micro-slips. Even very small amounts of wear can compromise positioning accuracy. Addressing these challenges is a central focus at ARCNL.

A key question is how industrially relevant multi-contact interfaces actually wear in practice. While nanoscale studies—such as transmission electron microscopy tracking the gradual wear of silicon tips or density functional theory (DFT)² calculations examining diamond/silicon oxide interactions—provide fundamental insights, they are difficult to translate to larger industrial interfaces. These large multi-contact systems are harder to probe experimentally, which leaves uncertainty about whether nanoscale mechanisms apply at that scale. Current work aims to bridge this gap, connecting nanoscale findings with the behavior of industrially relevant interfaces. Much of the focus is on stiff, wear-resistant materials such as silicon, which are widely used in semiconductor applications.

Repeated versus non-repeated sliding: surprising differences
Another important area of investigation is “non-repeated friction and wear.” Most tribology experiments use repeated motion—for example, pin-on-disc tests where the pin continually slides over the same track, or reciprocating tests with back-and-forth motion on the same region. These methods are useful for simulating systems, such as brakes or pistons, where repeated contacts dominate.

Some applications, however, involve non-repeated sliding. A familiar example is the wear of shoe soles: as someone walks, the sole does not repeatedly rub the same patch of floor but instead makes contact with new regions at every step. In experimental terms, this translates into sliding back and forth in one area, lifting the contact point and then repeating the cycle in a different area of the substrate.

At ARCNL, researchers conduct non-repeated friction and wear experiments using setups such as a ceramic ball mounted in a holder sliding across a silicon wafer. In these tests, the ball lands on the wafer, slides back and forth, lifts and then relocates to a new spot. This is an example of non-repeated ball-on-wafer testing.

Comparisons between repeated and non-repeated friction experiments show that the type of motion significantly affects the outcome. The mode of sliding—whether repeated or non-repeated—clearly matters. The key difference turns out to be what happens to the debris generated during sliding.

Debris buildup and its protective effect on surfaces
A comparison between repeated and non-repeated sliding experiments reveals striking differences in both friction and wear.

Two otherwise identical tests were conducted using the same materials, normal force, stroke length and velocity. The only distinction was motion type: in the repeated test, a sapphire ball slid back and forth over the same spot on a silicon wafer for 150 cycles, while in the non-repeated test, each cycle occurred on a fresh, untouched area of the wafer.

The outcomes differed significantly. In the repeated experiment, friction quickly stabilized at a constant level after a short run-in. In contrast, the non-repeated experiment showed a continuous increase in friction over time. Additional trials with other material pairs confirmed this behavior. For example, silicon carbide-on-silicon displayed the same trend—stable friction under repeated sliding but steadily increasing friction under non-repeated sliding. Even glass-on-glass experiments produced similar results.

Wear behavior also differed sharply. After repeated sliding, a layer of debris accumulated on the ball surface, forming a protective third body that reduced permanent material loss. In non-repeated sliding, the absence of this protective buildup led to significantly greater wear.

Quantitative analysis showed that non-repeated experiments resulted in seven times more wear than repeated experiments, despite all other parameters being identical. These findings highlight the importance of experimental design: When simulating applications that involve non-repeated sliding, it is critical to account for these differences, as repeated tests may underestimate true wear rates.

These findings prompted deeper investigation into the microscopic and even atomic-scale processes at play.

Atomic-scale wear: rare but powerful events
Closer examination of the wear behavior in non-repeated experiments revealed important insights. Optical microscopy of sapphire balls after testing shows smooth wear scars with defined diameters. Atomic force microscopy further indicates that within these scars, surface roughness is drastically reduced when compared to unworn regions. The roughness inside the scar measures around a nanometer or less—comparable to the roughness of the silicon wafer against which the ball was sliding. This suggests that the wear process may be driven by atomic-scale removal of sapphire.

A rough calculation supports this view. The real contact area at the ball–wafer interface is constrained by the hardness of the softer material—in this case, the wafer. By dividing the applied normal force by the wafer’s hardness, the minimum contact area can be estimated. Assuming a constant wear rate, the calculation takes the ratio of the total atoms removed to the atoms initially in contact, then divides the sliding distance by this number. Results show that to shave off just one angstrom³ of material, the system must undergo a far larger lateral displacement—implying that wear occurs as a rare event at the atomic scale.

While this calculation does not definitively prove atomic-scale wear, it strongly suggests that such a mechanism may be active in these experiments. Further studies on related systems provide additional evidence in support of this interpretation.

Beyond mechanics, surface chemistry also plays a critical role in how friction develops.

The role of surface chemistry in silicon-on-silicon friction
Recent research suggests that atomic-scale processes, such as covalent bond formation at interfaces, play an important role in friction behavior. In collaboration with the University of Amsterdam, experiments investigated silicon-on-silicon contacts using silicon balls with surface roughness of 20-40 nm against wafers that were much smoother, with roughness of about one nanometer.

A key observation was that plasma cleaning the silicon surfaces immediately before testing produced significantly higher friction coefficients when compared to untreated surfaces. The measured friction forces—on the order of 10s of mN⁴—scaled linearly with normal force, but the coefficient of friction increased by roughly a factor of three after plasma cleaning. This increase correlated with a major change in surface wettability: plasma-cleaned silicon became much more hydrophilic, as indicated by a lower water contact angle. So, surface chemistry clearly influenced the friction response.

The experiments also revealed a time-dependent effect. In a dry nitrogen environment, the initially high friction forces observed after plasma cleaning gradually decreased as the surfaces aged. At the same time, the water contact angle, which started out low, recovered to the higher values typical of untreated silicon. This mirrored evolution of friction and wettability suggested a common underlying mechanism.

When the same tests were conducted at 40% relative humidity, however, both the friction force and the contact angle remained stable for hours. Under these conditions, no decrease in friction or increase in contact angle was observed. The contrast between the two environments pointed to the role of surface chemistry and contamination.

One interpretation is that plasma cleaning produces hydroxyl-rich silicon surfaces that are highly reactive. In a dry nitrogen environment, trace carbonaceous contamination can gradually accumulate on these surfaces, reducing hydrophilicity and lowering friction. In a humid environment, by contrast, the presence of a water film prevents contamination buildup, maintaining the hydroxyl-terminated state. In this state, opposing hydroxyl groups on the silicon surfaces can form siloxane bonds across the interface, releasing water molecules and generating higher friction.

This chemical sensitivity pointed to an even more intriguing mechanism: the formation of covalent bonds across interfaces.

Evidence of covalent bonding at interfaces
To investigate the role of hydroxyl groups in friction behavior, researchers quantified hydroxyl density on silicon surfaces. A Cassie–Baxter model⁵ was applied, treating the surface as a combination of two fractions with distinct contact angles. At one extreme, freshly plasma-cleaned silicon surfaces were assumed to be fully hydroxylated, with a density of about four hydroxyl groups per square nanometer. At the other extreme, hydroxyl-free surfaces were assumed, corresponding to higher contact angles. By applying this model, measured water contact angles could be translated into hydroxyl group densities.

An experimental validation approach was also developed. Surfaces were treated with APTES⁶ and Rhodamine,⁷ which selectively bond to hydroxyl groups. Fluorescence microscopy was then used to detect the presence of Rhodamine molecules. Immediately after plasma cleaning, high fluorescence intensity was observed, indicating a high density of hydroxyl groups available for bonding. After six hours of drying, fluorescence intensity was significantly lower, suggesting a reduced hydroxyl density. Results from both the Cassie–Baxter model and the fluorescence labeling method provided consistent estimates of hydroxyl group density on the surfaces. These findings supported the hypothesis that friction in the system scales with hydroxyl density.

The probability of siloxane bond formation across the interface is proportional to the hydroxyl density on both contacting surfaces. Experiments confirmed this expectation: friction forces, normalized by contact area to calculate shear stress, aligned linearly with the predicted density of interfacial siloxane bonds. The slope of this relationship, approximately 0.6 nN per bond, corresponded closely to the known force required to break a siloxane bond.

Velocity-dependent friction tests provided further evidence. When surfaces were freshly plasma-cleaned, friction decreased strongly with increasing sliding velocity—a signature of bond-controlled friction, as faster motion allows less time for covalent bonds to form. This velocity weakening diminished after surfaces aged in nitrogen, or when a hydrophobic coating was applied, both of which reduce hydroxyl activity.

Together, these results strongly suggest that covalent bond formation at the interface plays a central role in friction between silicon surfaces. Such atomic-scale processes, occurring even at larger interfaces, represent an important mechanism that may also influence wear.

To test whether this atomic-scale bonding concept extended to other materials, researchers turned to diamond-coated spheres.

Diamond wear studies: angstrom-level insights
Recent work has focused on the wear of diamond coatings, studied using silicon carbide spheres coated with microcrystalline diamond and slid against silicon nitride wafers under controlled conditions. The experiments were conducted with a non-repeated methodology—each stroke taking place on a fresh section of the wafer—under both dry and ambient (50% relative humidity) environments.

Atomic force microscopy revealed gradual changes in the diamond surface over hundreds of cycles. Individual crystallites on the coating wore down and the silicon nitride wafers displayed scratches accompanied by material pile-up along their edges. Importantly, the depth and sharpness of scratches decreased over time, becoming almost undetectable by the 800th cycle.

By analyzing cross-sections of these scratches, researchers could estimate how much diamond material was lost per stroke. The results showed that wear proceeded at an atomistic scale—about half to one angstrom of diamond removed per cycle, despite sliding distances of 20 micrometers. This extremely slow rate aligns with the idea that wear occurs through rare atomic-scale events.

Environmental conditions proved critical. In dry nitrogen, the wear rate was roughly twice that observed under ambient humidity. This difference is consistent with the protective role of passivation species such as oxygen and water, which can adsorb to the diamond surface in humid environments and suppress atom removal.

Together, these findings reinforce the concept that even at macroscopic interfaces, wear of hard materials, such as diamond, can be governed by atomistic processes, strongly influenced by environmental chemistry.

These atom-by-atom insights then raised another issue: how surface roughness influences friction behavior.

How roughness influences slipperiness
An important question in tribology is why rougher surfaces can sometimes be more slippery. Traditional thinking holds that friction scales with the real area of contact, which in turn depends on surface roughness and applied load. In adhesive systems, such as polymers sliding on glass, this relationship is clear: larger contact areas lead to higher friction. However, other studies have shown that friction can instead scale directly with load, independent of contact area.

To investigate which regime applies to ceramic interfaces, researchers studied silicon nitride against sapphire under controlled conditions. Using both modeling and experimental visualization, they confirmed that these contacts behave elastically: smoother surfaces generate larger contact areas for the same load, while rougher surfaces produce smaller areas with higher contact pressures.

Friction experiments revealed that, regardless of roughness, friction force scaled linearly with load, yielding nearly identical coefficients of friction across samples. This indicated that the system followed load-controlled friction rather than adhesion-controlled friction. Still, a closer look uncovered a subtle but systematic effect: rougher surfaces produced slightly lower friction coefficients.

Modeling showed that this reduction could be explained by nanoscale capillary adhesion. On smoother surfaces, small water bridges form at the interface, effectively adding to the load and increasing friction. As surfaces become rougher, fewer such bridges form, lowering adhesion and reducing friction. Immersing the system in water eliminated the effect entirely, since capillary bridges could no longer form. These findings illustrate how, even in macroscale contacts, friction behavior is strongly influenced by nanoscale mechanisms. Small-scale adhesive effects, such as capillary bridges, can alter friction in ways not predicted by contact area alone, helping explain why rougher surfaces sometimes feel more slippery.

Together, these discoveries form a comprehensive picture of friction and wear at multiple scales.

The study of friction and wear in multi-contact interfaces reveals that even at macroscopic scales, atomistic processes—such as single-layer wear events and covalent bond formation—play a critical role in determining system behavior. By bridging the gap between fundamental science and industrial application, this research provides semiconductor manufacturers with key insights into managing nanoscale interactions that directly affect wafer handling, alignment and equipment longevity. These findings not only advance the scientific understanding of tribology, but also help ensure the precision, reliability and efficiency required for the next generation of semiconductor devices.

Acknowledgements
Feng-Chun Hsia
Fiona Elam
Liang Peng
Cyrian Leriche
Pierre Audebert
Daniel Bonn
Chen Xiao
Chao Hsu
Steve Franklin
Fred Brouwer

Presentation resources
• Hsia, F.-C., Elam, F.M., Bonn, D. Weber, B. and Franklin, S. E. (2020), “Wear particle dynamics drive the difference between repeated and non-repeated reciprocated sliding,” Tribol.Int., 142, 105983, pp. 1-8.
• Peng, L., Hsu, C. C., Xiao, C., Bonn, D. and Weber, B. (2023), “Controlling macroscopic friction through interfacial siloxane bonding,” Phys Rev Lett., 131 (22), 226201. DOI: 10.1103/PhysRevLett.131.226201. PMID: 38101386.
• Leriche, C., Pedretti, E., Sahin, O., Kang, D., Righi, M.C. and Weber, B. (2025), “Passivation species suppress atom-by-atom wear of microcrystalline diamond,” ACS Appl. Mater. Interfaces, 17, 39, 55511-55520, https://doi.org/10.1021/acsami.5c08647.
• Hsia, F.-C., Franklin, S., Audebert, P., Brouwer, A. M., Bonn, D. and Weber, B. (2021), “Rougher is more slippery: How adhesive friction decreases with increasing surface roughness due to the suppression of capillary adhesion,” Phys. Rev. Res., 3, 043204. DOI: https://doi.org/10.1103/PhysRevResearch.3.043204.
Notes
1. Public partners are the University of Amsterdam, the Free University of Amsterdam, the Dutch Science Organization NWO and Groningen University. The collaborative industrial partner is semiconductor equipment manufacturer ASML.
2. Density functional theory is a quantum mechanical modeling method used to calculate the electronic structure of atoms, molecules and solids based on electron density rather than wavefunctions.
3. An angstrom (Å) is a unit of length equal to one ten-billionth of a meter (10⁻¹⁰ m), commonly used to measure atomic and molecular scales.
4. A milinewton (nN) is one-thousandth of a newton.
5. The Cassie–Baxter model describes how a liquid droplet rests on a rough or textured surface, where air pockets trapped beneath the droplet reduce solid–liquid contact and increase water repellency.
6. APTES (3-aminopropyltriethoxysilane) is a silane coupling agent used to functionalize surfaces with amine groups.
7. Rhodamine is a fluorescent dye commonly used for imaging and labeling in biological and material sciences.

Jeanna Van Rensselar heads her own communication/public relations firm, Smart PR Communications, in Naperville, Ill. You can reach her at jeanna@smartprcommunications.com.