---

---

KEY CONCEPTS
• Critical shortcomings such as limited thermal stability and strong dependence on environmental conditions for optimal performance have inhibited the widespread use of solid lubricants. Reducing the effect of these deficits can be achieved through a deeper understanding of the molecular mechanism that directs the friction and wear response. A key step is the identification of chemical and structural changes at sliding interfaces.
• Evaluation tools include proven methods such as atomic force microscopy and transmission electron microscopy and, more recently, computerized simulation and modeling.
• By combining methods, it has been possible to solve such puzzles as the origin of super-low friction of diamond-like carbon coatings combined with organic friction modifiers and of their high wear rates when they are used in the presence of ZDDP.

In most tribological applications, liquid or grease lubricants are adequate for reducing friction and wear. However, when operating conditions become very severe, as in extreme temperatures or in the presence of ionizing radiation, solid lubricants may be the only choice for reducing friction in mechanical systems.

Solid lubricants most often are used where liquid lubricants won’t work—this can be due to operational reasons such as a wide temperature range (viscosity and/or degradation issues for liquid lubricants) or the need for vacuum or clean environment (vapor pressure issues for liquid lubricants).

Dr. Julien Fontaine, CNRS (Centre National de la Recherche Scientifique, French National Centre for Scientific Research) research scientist, working at the Laboratory of Tribology and Dynamics of Systems (LTDS) at  École Centrale de Lyon, explains, “This is why solid lubricants are used for the outer parts of satellites, that may slide at relatively low speed, over a wide temperature range (-100 to 100 C) and under vacuum. Another famous example is related to diesel engines: the fuel injectors are using diamond-like carbon (DLC) coated needles to prevent wear and seizure in the needle tip/injection hole contact as well as to reduce wear and friction losses in the guiding part of the needle. The use of liquid lubricants is not possible, as they would be mixed and evacuated with the fuel itself—although the diesel fuel can be considered relatively lubricious.”

At sliding interfaces, solid lubricants should create a low shear-strength layer that reduces friction. Several inorganic materials such as molybdenum disulfide, hexagonal boron nitride and carbon-based materials (i.e., DLC and graphene) have been used in a range of environmentally demanding appli- to manufacturing because of their excellent lubricating properties.

While most of these solids owe their lubricity to a lamellar or layered structure (e.g., graphite, molybdenum disulfide), some others (e.g., DLC, polymers) do not have a layered crystal structure. For these non-lamellar solids, the ability to reduce friction is attributed to the shear-induced structural transformations and chemical reactions occurring in their near-surface regions. These solids create surface layers with inherently low shear strength that leads to friction reduction.

While solid lubricants have many proven benefits in a range of applications, critical shortcomings have inhibited their more widespread use. In particular, most solid lubricants exhibit limited thermal stability and strong dependence on environmental conditions for the scale of their tribological response. Liquid lubricants also have the function of taking the frictional heat out of the contact. Thus, at high sliding speeds, solid lubricants may endure larger temperatures than in a liquid-lubricated contact.

The design of novel solid lubricants with enhanced tribological behavior across a range of conditions and environments depends on the establishment of a deeper understanding of the molecular mechanism dictating the friction and wear response of these materials. A key step in the development of this understanding is the identification of the chemical and structural changes occurring at sliding interfaces. 

The identification of the material properties needed to achieve the desired tribological response will enable a customized approach and open the path for the discrete design of novel solid lubricants with improved friction and wear properties. The development of the required scientific knowledge to achieve this goal is predicated on fundamental and applied research, which encompasses both experiment and simulation. The use of a highly multidisciplinary, integrated, multi-technique approach is required because of the complex phenomena occurring at sliding interfaces; heat, diffusion, phase transformations and surface and interfacial reactions all combine at tribological interfaces to produce highly unstable conditions. The multidisciplinary issues emerging from these situations demand high-quality research in solid lubrication.

In the last decade, the knowledge base for solid lubrication has been advancing from empirically based to scientifically based because of the introduction of novel analytical methods with improved sensitivity and resolution as well as the development of powerful computational methods.

Addressing limitations
DLC films, in particular, are highly susceptible to their operating environment. Vacuum and dry conditions exacerbate the dichotomy between these two extreme behaviors. Under vacuum, highly hydrogenated films may provide very low (or even super-low, μ < 0.01) coefficients of friction, while less hydrogenated films will exhibit very high (μ > 0.7) friction coefficients and wear. In more humid environments, such as ambient air, the coefficients of friction will be less extreme, in the range 0.1-0.25.

Fontaine observes that another limitation is sliding speed and its effect on temperature: Since the DLC structure is amorphous and meta-stable, high sliding speeds may lead to thermal heating of the rubbed surfaces. High temperatures (above 300 C) will affect the structure of the DLC, favoring the release of hydrogen and the formation of C–C bonds (by re-hybridization from sp³ C to sp² C). Moreover, in oxidizing environments, the DLC surface also may be etched, increasing the wear and affecting the friction response.

“Some attempts have been made with doped DLC to address these issues, in order to stabilize the amorphous meta-stable carbon network and protect it from oxidation,” Fontaine says. “Silicon or silicon oxide-doped DLC, for instance, may provide some improvement, but the oxidation residues are, thus, silica-based, with a tribological behavior that may not be as good. Maybe other doping agents could be more efficient? This is still under investigation.”^1,2

Another approach Fontaine thinks is worth considering is nanocomposite materials, with different phases combined together, each addressing a specific limitation (for instance, different environments). “Such materials are sometimes called chameleon coatings,” he says. “The tribological performances are not as extreme as pure solid lubricants, but they are less sensitive to operating conditions.”

Molecular-level mechanisms of solid lubricants
While there is not a single lubrication mechanism of DLC because of the many different materials and properties as well as numerous environments and operating conditions, there are some general trends.

“First, it seems that adhesive phenomena govern the tribological behavior of DLC films (providing the sliding surfaces are not too rough, of course),” Fontaine explains. “When adhesive junctions are formed, they have to be broken to allow sliding, and the way these junctions are broken is, thus, paramount. Depending on the nature of the counterface as well as of the environment, the breaking may take place within the DLC, leading to a tribofilm buildup on the counterface.”³

Fontaine says that in his common work with STLE member Filippo Mangolini and others, they have learned that the environment helps to break in the DLC. “This is probably due to chemical interactions,” Fontaine points out. “The release should be controlled by mechanical properties and take place in the weaker material, either softer or more brittle, but stress-induced chemical reactions may help open and propagate localized cracks in the DLC side of the adhesive junction.”

Another key parameter controlling the release of adhesive junctions is their size and distribution; a single large junction would be more difficult to break (and would generate more wear) than many small ones. A parallel Fontaine uses is that a pinned sheet of paper breaks more easily than one that is not pinned. Once adhesive junctions have been broken, the transferred material as well as the nascent surface of the DLC may be modified quite significantly.

“The structure of DLC films is amorphous and meta-stable, thus such breaking may trigger structural evolutions,” he explains. “Also, broken C–C bonds need to be made passivated—either these bonds will become passive by reacting with molecules from the environment or by re-hybridization with their carbon neighbors. The triggering between both will depend on the nature and pressure of the molecular environment, as well as the exposure time—re-hybridization should be favored by low gas pressure and/or high sliding speeds. Anyway, the modification on the DLC surface and tribofilm may affect the further evolution of adhesion between the sliding surfaces, and, thus, of the tribological response. Finally, the solid lubrication processes of DLC films are strongly dependent on the mechanics, chemistry and nanoscale roughness of the contacting surfaces.”

Alloying elements
Tribological properties of DLC films can be improved using alloying elements in different ways. The amorphous structure is a real advantage because a wide range of elements can be incorporated without being constrained by compatibility with a crystalline matrix.

STLE member J. Brandon McClimon, postdoctoral researcher, University of Pennsylvania, Department of Mechanical Engineering and Applied Mechanics, notes, “Alloying goes back a long way, with the usefulness of hydrogen as a doping or alloying element recognized decades ago and quickly incorporated into technical applications like hard drive overcoats. Hydrogen provides surface passivation and generally allows for very low friction and wear in inert environments. Fluorine provides similar advantages and reduces the surface energy of the DLC, although wear resistance tends to be somewhat poorer.”

McClimon says that his own research has focused on silicon oxide doping, which is easier and safer than pure silicon doping since there is no need to use pyrophoric silane gas during the deposition and it provides similar functionality. The primary advantage is a significant improvement in thermo-oxidative stability and improvements in pure thermal stability as well. “There are many additional elements that can be used whose advantages are empirically known, but I think it is fair to say some of the detailed physical mechanisms are still being unraveled by groups all over the world,” he adds. “These include elements like sulfur which can, like silicon, reduce the humidity dependence of friction and wear; metals like titanium and chromium, which improve adhesion to the substrate; and perhaps related, the internal stress state of the deposited films. Noble metals and copper also can be incorporated to improve electrical conductivity, and there are more.”



Choice of the alloying element
McClimon explains that the choice depends on the response to be modified. For example:
• If better thermo-oxidative stability is desired, silicon or silicon oxide is a good choice.
• If better electrical conductivity is desired, copper or one of the noble metals might be a good choice.
• If an improvement in fracture toughness or adhesion to the substrate to allow for thicker films is desired, titanium or chromium metal doping are good choices.

“The choice to alloy must always be balanced against inevitable tradeoffs,” he says. “For instance, in nearly all cases, alloying results in some reduction of the sp3 fraction and an associated reduction in the wear  resistance. In many cases, the overall functional response still is improved, though.”

Lubrication mechanisms of alloyed solid lubricants
McClimon thinks that, in general, the lubrication mechanisms of alloyed DLCs, to the degree they are understood, largely match those of unalloyed DLCs. “Surface passivation with species that prefer a single chemical bond, typically hydrogen or fluorine, is critical across all DLCs,” he says. “A surface terminated purely with carbon will almost inevitably bond with the counterface of the tribocontact, leading to high friction and wear. The passivating species can be provided by the ambient environment or doped into the film itself. Simulations also show that the restructuring of the DLC film to orient bonding directions parallel to the sliding interface is an important mechanism for reducing the friction.”⁴

The role of the counterface, and specifically, the tribofilm that transfers from the DLC to the counterface, increasingly has become the focus of research. It appears that tribofilms are required for low friction in most DLC sliding systems, with the exception of perhaps some superlubricious DLCs highly doped with hydrogen, which can attain low friction without any obvious evidence of tribofilm formation.

“My own work shows that even well-passivated DLC-on-DLC sliding interfaces can show extremely high shear strengths in the absence of a tribofilm even with no evidence of interfacial chemical bonding,” McClimon adds.⁵

He continues, “How DLC tribofilms provide lubricity is not fully explained yet, but useful new context has been published frequently in recent years. Surface passivation of the tribofilm itself is important, with clear evidence that the near-surface layer of tribofilms is enriched with hydrogen. Interfacial adhesion is substantially reduced by tribofilms, which reduces the real area of contact and, therefore, the friction. Tribofilms also are usually much softer and less stiff than the starting DLC, which could make de-pinning of locked nano-asperities easier and reduce the friction—although evidence for this mechanism is thin currently.”

On the subject of evaluating the lubrication mechanisms of solid lubricants, McClimon summarizes, “All of that said, alloying elements aside from hydrogen can modulate the friction response in important ways. For instance, silicon oxide and sulfur can reduce the typically strong dependence of friction on the environmental humidity. In the case of silicon oxide, we have a proposed mechanism for this, which is that the favored hydroxyl termination of silicon attracts and tightly binds water molecules, which subsequently serve as a boundary lubricant.”⁶

Methods of evaluating the lubrication mechanism
Regarding the evaluation of DLC films, McClimon is partial to atomic force microscopy (AFM) since that is what he most frequently uses. The atomic force microscope allows for measuring friction forces down to a single nanoscale asperity contact. This simplifies the geometry relative to macroscale sliding since the real area of contact can be estimated using simple contact mechanics models and extracting critical quantities, such as the interfacial shear strength.

AFM does not, however, allow one to see the chemistry and structural changes occurring at the sliding interface. “This is where advanced spectroscopies come in,” McClimon says. “X-ray photoelectron spectroscopy, scanning Auger electron spectroscopy and secondary ion mass spectroscopy can provide accurate composition and some chemical information about the surface of the sample. Synchrotron-based X-ray absorption spectroscopies can provide much more detailed chemical information, about carbon bonding states in particular, and often have better lateral resolution, which can be useful for characterizing AFM tip-modified areas, which are inevitably nanoscale in extent.”

McClimon continues, “Transmission electron microscopy (TEM)-based electron energy loss spectroscopy is extremely powerful and has been the focus of some very good recent papers on DLC tribology. The advantage here is the extreme lateral resolution, sub-nm in the case of aberration-corrected TEMs, along with the ability to look below the surface with focused ion beam (FIB) milling of cross-sectional samples. In particular, tribofilms present a very challenging geometry for spectroscopic characterization that is quite tractable in the case of FIB-milled cross-sectional TEM samples.”⁷

While there are numerous evaluation methods that remain important because of their advantages, in McClimon’s view, the surface force apparatus provides the best quantification of the real area of contact in low normal stress contacts. “Nanoindentation remains the best way to characterize the mechanical properties of these thin film samples,” he says. “Even the standard pinon-disk tribometer, when mounted inside ultrahigh vacuum, is a powerful tool for understanding the influence of environmental species on the friction and wear response of DLCs, which is profound across all DLCs.”

Computational/simulation modeling of the lubrication mechanism
Computational modeling involves the input of mathematical and physics variables into a computer with the goal of simulating a complex system. Simulation is accomplished by modifying and recombining the variables in order to find a numerical solution of the mathematical models that describe the behavior of the complex system.

The field of research for Dr. Gianpietro Moras, Fraunhofer Institute for Mechanics of Materials IWM and MicroTribology Center μTC, Freiburg, Germany, is focused on computational methods with atomic-scale resolution (i.e., atomistic simulations).

“In practice, using these computational methods, we can use supercomputers to calculate how atoms interact with each other when a tribological load is imposed on them and follow their trajectories over time,” he explains. “This is a very powerful tool to complement experimental analyses and look into the details of the physical and chemical processes taking place at tribological contacts. These details are otherwise virtually inaccessible to even the most advanced experimental techniques.”

Friction and lubrication are the result of atomic-scale interactions between the surfaces that form the tribological contact (i.e., interacting surfaces in relative motion) and between these surfaces and the molecules of the lubricant or the surrounding environment—humid air, for instance. Moras says, “Using atomistic computational methods, we can gain insights into these interactions and into the chemical and structural transformations the tribological system undergoes when a tribological load is applied. In this way, we can identify friction mechanisms with atomic-scale resolution, or we can verify hypothetical mechanisms brought forward through colleagues’ experiments.”

Once a possible mechanism is identified, atomistic computational methods can be further used to perform high-throughput screening and parametric studies. Moras adds, “For instance, we can perform targeted chemical modifications of the systems (such as alternative lubricant formulations or chemical modifications of the thin films), or we can change the mechanical boundary conditions (e.g., contact pressure) and investigate how these modifications affect the friction mechanisms. This provides useful insights for the design and optimization of tribological systems.”

Moras gives the following example of how this knowledge can help in the understanding of the lubrication mechanisms of solid thin films: Consider two surfaces that are coated with very smooth thin films made of diamond or DLC. Experiments show that, under some circumstances (e.g., specific mechanical loads and sliding speed combined with the right friction modifier), these tribological systems can exhibit super-low friction (i.e., a friction coefficient lower than 0.01). While the friction properties can be easily measured experimentally, even the most accurate chemical characterization of the tribological surfaces (which is necessarily performed only before and after friction occurs, not during sliding) cannot unequivocally reveal the origins of superlubricity (see Misleading Terminology). Using all the input that can come from these experimental analyses, an atomic-scale model system could be set up that is as similar as possible to the experimental one, and the simulations could be used as a digital magnifying glass to follow the processes that take place during friction.⁸ 

---

Misleading terminology
Dr. Julien Fontaine, CNRS (Centre National de la Recherche Scientifique, French National Centre for Scientific Research) research scientist, working at the Laboratory of Tribology and Dynamics of Systems (LTDS), has issues with the term “superlubricity.” “I personally would be very cautious using this word, because it is originating from an analogy with superconductivity,” he explains. “If superconductivity is the vanishing of electrical resistance, superlubricity would be a vanishing of resistance to sliding, but there is a major difference. Superconductivity is a material property, while the achievement of super-low friction is a contact response, which depends on the contacting materials, the operating conditions, the environment and so on. So this cannot be seen as a material property, and the word ‘superlubricity’ is, thus, in my opinion, misleading.”

---

Moras poses three questions that simulation can answer.
1. How does the atomic-scale structure of the carbon films evolve during sliding? “Our simulations show that even initially crystalline surfaces, like those of diamond films, develop an amorphous surface layer that is significantly different from the surface of the films prior to their tribological loading,” Moras says.
2. How do the surfaces of the thin films interact between them or with the friction modifier, and why does this particular friction modifier under these circumstances lead to super-low friction? Moras continues, “Once again, the simulations prove useful and reveal the following scenario: If no molecules are present between the two surfaces (i.e., in vacuum), strong chemical bonds form between the two reactive surfaces. Moving one surface relative to the other requires large forces because these bonds must be broken. This results in high friction coefficients that are measured experimentally.”
3. Are there ways to optimize this process by controlling the chemistry of the surfaces or of the friction modifiers? Moras explains that simulations can provide useful guidelines for design here, too. For instance:
• Suitable chemical modification of the carbon film surfaces can favor their chemical passivation by replacing surface “dangling bonds” with hydrogen atoms or hydroxyl groups or by promoting surface aromaticity (a kind of graphitization). This prevents the formation of covalent bonds between the two sliding surfaces, thus, leading to low friction.
• Ideal friction modifiers are organic molecules that have two or more reactive sites, like oleic acid or glycerol, so that the molecules can simultaneously attach to both sliding surfaces, facilitating their fragmentation and the release of oxygen and hydrogen atoms that chemically modify the DLC surfaces favoring their passivation.

Another of Moras’ examples⁹ involves an analogous system in which surfaces are coated with DLC thin films: ZDDP is one of the most indispensable and effective antiwear additives for automotive combustion engines. However, while it works very well for ferrous materials, it can seriously affect the friction and wear reduction of DLC in automotive applications. By combining experiments, contact mechanics simulations and atomistic modeling, it has been possible to unveil the origin of high wear rates in DLC coatings that are used in the presence of ZDDP. It turns out that high wear rates are caused by mechanically induced ZDDP fragmentation and sulfur release, which weakens the carbon film. This enabled the investigation into whether this mechanism can be hindered and led to the discovery that high wear rates can be avoided by properly tailoring the DLC coating’s stiffness, surface nano-topography and hydrogen content.

“Simulations typically provide the clearest identification of atomic-scale mechanisms driving empirical responses observed in experiments,” McClimon adds. “Realistic films can be deposited inside of the simulations including alloying elements. These computational depositions include important physics like the subplantation mechanism by which incident ions elevate the sp³ fraction of the films, so the resulting films are thought to be a good match for experimental films, structurally.”



Limitations of simulation
As with most simulation applications, cost is a major factor. While the price of both on-premises computer hardware and cloud computing is high, it’s often the software that represents the greatest expense. This puts simulation beyond the reach of many smaller enterprises and start-ups. “The main limitation of atomistic computational methods whose chemical accuracy is high enough to be useful in tribology is their computational cost,” Moras explains.

Other indirect cost deterrents Moras cites are:
• Time. Typically useful simulation studies require months of simulations on large supercomputers.
• Expertise. Setting up meaningful simulations and interpreting the results require years of experience in the field.

“Typically, depending on the level of accuracy that is required to suitably model the problem, the length/time scales that can be simulated using state-of-the-art high-performance computers is limited to about 1 nanometer/1 nanosecond for quantum mechanical methods that can accurately describe the chemistry of systems with arbitrary chemical elements, while it is limited to about 100 nanometers/100 nanoseconds for classical mo- lecular dynamics that can describe the structural evolution of systems with one or two chemical elements (e.g., carbon and hydrogen),” Moras explains, summarizing these limitations:
• Sliding speeds are usually much larger than typical speeds used in technological applications.
• The system size is limited, and only roughness on the nanometer scale can be modeled directly.
• Increasing the size and simulation time to about 100 nanometers/100 nanoseconds can only be achieved for systems that are chemically relatively simple, which severely limits the amount of technologically relevant systems that can be simulated at these scales.
• Owing to the limitations described in the previous points, setting up a meaningful model system heavily relies on experiments that ideally have to provide accurate information about the nanoscale chemical structure and mechanical boundary conditions.

“So, simulating the deposition of alloyed DLC films to determine their structure is limited to very thin films, which might not be representative of the much thicker experimental films,” McClimon says. “Simulations of tribological response are limited to very short time periods, typically nanoseconds, so there is not time to grow significant tribofilms that profoundly modulate the friction and wear response. The short durations typically also necessitate very high sliding speeds to ensure enough relative motion so that relevant phenomena occur.”

That said, McClimon does believe that a combination of experiments and simulations can often allow for identifying structures and mechanisms important to understanding the attributes of alloyed DLC films.

Moras concurs: “Although there is still some way to go until simulation-based design of solid lubricants becomes routine, there are many successful examples of the rational design of solid lubricants in which simulations played an important role. Of course, these examples are rarely the result of simulations alone or of simulations involving a single modeling scale.”

How these limitations can be addressed Moras explains that there are three possible approaches to address these limitations that should be adopted in conjunction with each other. The first two rely on the simultaneous use of a variety of existing modeling and experimental methods. The third approach involves the development of novel computational techniques with improved efficiency without compromising the accuracy.

The first approach is to extend modeling predictions to larger length scales and longer time scales by connecting several modeling methods at different scales (multiscale modeling). This is based on a flux of information that runs back and forth across scales. For instance, coarse graining techniques gather information at the smaller, more accurate scale to feed coarser models.

“This can be done by using the results of quantum-mechanical simulations to formulate kinetic models based on reaction rates—or to develop coarser molecular dynamics methods that are no longer able to describe chemical reactions but can describe physical interactions when the system under investigation does not undergo relevant chemical transformations,” Moras clarifies. “In the other direction, the mechanical boundary conditions (e.g., nanoscale contact pressure) can be extracted from larger-scale contact mechanics simulations that are based on contact mechanics.”

The second approach Moras endorses involves a very close link between simulation and experiments. “The more detailed the information we receive from our experimental colleagues about chemical structure, surface roughness or even hypothetical friction mechanisms, the easier it is to set up meaningful model systems that contain the essential ingredients and can be simulated with the available computational power,” he says. “While these complex experimental analyses are often impractical in technological applications, this step requires a significant effort by our experimental colleagues who have to design auxiliary experiments that have to be performed on carefully controlled systems (e.g., in terms of chemistry and surface roughness) and carefully controlled environments (e.g., lubricant composition).”

The third way to address the limitations of computational modeling, according to Moras, is the development of more efficient modeling techniques that do not compromise accuracy—an effort that is currently benefiting from modern developments in machine learning.

New experimental methods for development
McClimon believes that, while speculative, the further development of in situ TEM techniques is very promising for the study of DLCs. “Recently, the ability to image samples in ambient environmental pressure has been developed, which is hugely advantageous for a variety of scientific areas,” he says. “At the same time, the ability to perform sliding experiments with atomic force microscope probes inside the TEM also has been developed, allowing for quantification of various tribological response parameters while observing the atomic-scale mechanisms causing them, albeit in an ultra-high vacuum environment.”

He adds that if the ability to combine sliding functionality under the TEM with relevant ambient environments was realized, this would be transformational in the following ways:
• Interactions of alloying elements with both the bare and tribofilm-covered counterface could be examined.
• The transfer mechanisms of the alloying elements to the counterface and subsequent incorporation into the tribofilm could be tracked in detail over time.
• Nanometer-resolution electron energy loss spectroscopes could determine local stoichiometry, bonding state and density evolution throughout the experiment with the effect of alloying elements determined locally in both the DLC and the tribofilm.

“Such further development would be transformative for our ability to understand alloyed DLC tribological response in relevant environments,” McClimon observes.

Areas for further research
According to McClimon, while many of the basic tribological mechanisms of solid lubricants are understood, there are research areas that still need to be addressed. He points to the need to research thicker films that would extend the wear life of components in some applications.

“Fundamental tribological questions about DLC lubrication and wear still remain that also could prove important for further technological development,” he says. “The fundamental mechanism for tribofilm lubricity still remains unknown. For the DLC side of the interface, simulations have proved incredibly useful in identifying many processes relevant to understanding DLC lubricity. The tribofilm side of the interface cannot be simulated accurately at present because the structure of the tribofilm is not known in sufficient detail. There also is the important question of whether DLC superlubricity can be reliably extended to ambient environments. The wide adoption of DLC films with superlubric friction could save enormous amounts of energy. Alloyed DLCs have a role to play here, as the existing demonstrations of superlubric DLC contacts in humid air were achieved with elemental doping, specifically sulfur, silicon and silicon/titanium doping. However, the fundamental reasons superlubricity was achieved are still under debate, and films of similar compositions and structures do not always show superlubricity in testing.”

Collaboration toward better design 
Common activities performed by computational scientists to generate new ideas can significantly speed up the development process. Moras says success requires open communication and an interdisciplinary approach in the following manner:
1. An effective and open communication channel must be established between industry and research actors to bridge the communication gap between industry and research and understand what the needs and problems of industry are.
2. Computational and experimental researchers need to team up, sit together and plan/design both experimental and computational activities, which go hand in hand.
3. Computational researchers have to focus more on problem solving and less on the particularity of their own simulation method. They need to master modeling methods at different scales and physical levels (e.g., quantum mechanics, molecular dynamics, computational fluid mechanics, continuum mechanics) or team up with other computational researchers in order to be able to operate in a multi-scale fashion. Fontaine adds, “Collaborating with theoreticians should be fruitful, first to further understand experimental observations, and then to consider potentially new solid lubricants. The challenge for theoreticians working with numerical modeling is to find a way to address phenomena taking place at different scales, with chemical interactions that occur at atomic scale, while the volume of material involved in the mechanical response strongly depends on the contact size—usually in the micrometer range and above.

Moras concludes that by combining accurate experimental inputs and computer simulations, the main lubrication mechanisms of DLC materials can be identified and lead to an understanding of which chemical and mechanical modifications of the system can activate or suppress them.

Fontaine believes that the question of liquid versus solid lubricants is getting less relevant and that, for many applications, they will be used in combination—the solid lubricant acting as a backup when liquid lubrication may fail (low-speed/high-contact pressure, for instance).

As multidisciplinary research into solid lubricants continues, McClimon expects that the range of applications will expand significantly. “DLCs are quite widespread already,” he summarizes. “When I first started my doctorate, the only automotive applications were in NASCAR and the F1 circuit. Now you see DLC coatings on the piston rings of $20,000 cars and in many other automotive applications. Hydrogenated amorphous carbon (a-C:H) films have been used as hard drive overcoats for decades and have been used on tools for machining aluminum for many years. Razor blades have been coated with DLCs for a long time. Biomedical implants are a huge growth area for DLCs since the carbon-based materials are biocompatible. Manufacturing equipment, including equipment for semiconductor manufacturing, is another growth area for DLCs.”

REFERENCES
1. Koshigan, K. D., Mangolini, F., McClimon, J. B., Vacher, B., Bec, S. Carpick, R. W. and Fontaine, J. (2015), “Understanding the hydrogen and oxygen gas pressure dependence of the tribological properties of silicon oxide–doped hydrogenated amorphous carbon coatings,” Carbon, 93, pp. 851-860. Available here.
2. Mangolini, F., Koshigan, K. D., Van Benthem, M. H., Ohlhausen, J. A., McClimon, J. B., Hilbert, J., Fontaine, J. and Carpick, R. W. (2021), “How hydrogen and oxygen vapor affect the tribochemistry of silicon- and oxygen-containing hydrogenated amorphous carbon under low-friction conditions: A study combining x-ray absorption spectromicroscopy and data science methods,” ACS Applied Materials & Interfaces, 13 (10), pp. 12610-12621. Available here.
3. A tribofilm is formed from the DLC material but might be different in its structure and composition, thus the use of a specific name. On the other hand, a transfer film suggests a transfer but not a modification.
4. Pastewka, L., Moser, S. and Moseler, M. (2010), “Atomistic insights into the running-in, lubrication, and failure of hydrogenated diamond-like carbon coatings,” Tribol Lett., 39, pp. 49-61. Available here.
5. McClimon, J. B., Hilbert, J., Lukes, J. R. and Carpick, R. W. (2020), “Nanoscale run-in of silicon oxide-doped hydrogenated amorphous carbon: Dependence of interfacial shear strength on sliding length and humidity,” Tribol Lett., 68, 80. Available here.
6. Kajita, S. and Righi, M. C. (2016), “A fundamental mechanism for carbon-film lubricity identified by means of ab initio molecular dynamics,” Carbon, 103, pp. 193-199. Available here.
7. Chen, X., Zhang, C., Kato, T., Yang, X., Wu, S., Wang, R., Nosaka, M. and Luo, J. (2017), “Evolution of tribo-induced interfacial nanostructures governing superlubricity in a-C:H and a-C:H:Si films,” Nature Communications, 8, 1675. Available here.
8. Kuwahara, T., Romero, P. A., Makowski, S., Weihnacht, V., Moras, G. and Moseler, M. (2019), “Mechano-chemical decomposition of organic friction modifiers with multiple reactive centres induces superlubricity of ta-C,” Nature Communications, 10, 151. Available here.
9. Salinas Ruiz, V. R., Kuwahara, T., Galipaud, J., Masenelli-Varlot, K., Ben Hassine, M., H au, C., Stoll, M., Mayrhofer, L., Moras, G., Martin, J. M., Moseler, M. and de Barros Bouchet, M. I. (2021), “Interplay of mechanics and chemistry governs wear of diamond-like carbon coatings interacting with ZDDP-additivated lubricants,” Nature Communications, 12, 4550. Available here.

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