Low-Energy Triboelectrons and Tribochemistry - A ReviewC.K. Kajdas
Warsaw University. of Technology, Institute of Chemistry at Plock, 09-400, Plock, Poland, e-mail: ckajdas @ pw.plock.pl
AbstractA review is presented of both (i) triboemission of low-energy electrons and (ii) their importance for triggering the initiation of tribochemical reactions. Mechanical energy associated with the boundary lubrication process releases a large number of physical phenomena. Energy outputs from rubbing can be the cause of tribochemical reactions between the friction components and lubricant molecules. Experimental results, mostly on triboemission from diamond-on-ceramics and diamond-on-semiconductors are discussed. Emphasis is on the role of triboemitted charged-particles on the initiation of tribochemical reactions. The hypotheses and mechanisms regarding the possible action of charged-particles, especially low-energy electrons, on tribochemical processes are reviewed. Specific connections are made between relevant tribochemical reactions, tribopolymerization among them, the Negative-Ion Radical Action Mechanism (NIRAM) and recent experimental results in the fields of triboemission and tribochemistry. An important part of this review presentation considers these authors' experimental finding, that mostly low-energy electrons (0 - 5eV) are triboemitted from the investigated ceramics. The finding is discussed with a focus on the mechanism of vinyl-type tribopolymerization. A new instrument for the measurement of charge intensity and energy spectrum of triboemitted particles is also demonstrated. Additionally, experimental research work is presented seeking a better understanding of triboemission of negatively-charged particles from sliding contacts. This paper also reviews the most recent research on negative-charge intensity from the semiconductors silicon and germanium when they are scratched by a diamond pin in UHV. While significant triboemission was always detected during the sliding contact, emission was undetectable when contact ended. Emission intensity is substantially higher for diamond-on-Si than for diamond-on-Ge.
Key-words: triboemission, tribochemistry, tribopolymerization, ceramics, semiconductors, NIRAM
INTRODUCTIONTribochemical reactions are distinct from those of thermochemical ones. Mechanical energy associated with the boundary lubrication process releases a large number of physical phenomena. Energy outputs from rubbing can be the cause of tribochemical reactions between the friction components and lubricant molecules. Such reactions are often related to mechanochemistry. Triboemission is defined as the emission of electrons, ions, neutral particles, photons, radiation and acoustic emission under conditions of tribological damage. Triboemission of electrons is a particular case of the general phenomenon of exoelectron emission (EE). Although the mechanism for EE is not well understood, EE has been observed from both metals and non-metals and there is strong evidence that oxides or other non-metallic surface layers are needed for EE to occur. The negative-ion-radical action mechanism approach [1-3] assuming that tribochemical reactions are initiated by low-energy electrons in the energy range of 1 to 4 eV. Based on measurement of triboemitted charged particles from ceramics, first evidence for such energy range was presented by Molina et. al . Low-energy triboemitted electrons have been demonstrated to play a significant role in tribochemical reactions under boundary lubrication, an example of which is the mechanism of tribopolymerization, as presented by Furey et al. . Selected addition-type monomers polymerizing by an anionic mechanism form anti-wear tribopolymers on the presence of low-energy electrons. The importance of low-energy triboelectrons for the initiation of tribochemical reactions most recently was discussed by Molina et al. . It was concluded that by knowing more about electrons emitted during sliding and wear, it could be possible to match monomers with systems in a rational approach, rather than a trial-and-error, to boundary lubrication via tribochemistry. Another example may relate to the degradation mechanism of perfluoropolyethers (PFPE) with DLC coatings in thin film magnetic rigid disks . It was concluded that anionic intermediates (negative ions and/or negative-ion-radical species) produced by low-energy electrons play an important part in both (a) the electron mediated degradation process of PFPE lubricants, and (b) chemical bonding of PFPE lubricant films with DLC surfaces under sliding conditions. Accordingly, it is believed that the dominant degradation mechanism of PFPE lubricants is the low-energy electron degradation mechanism .
Looking at other evidences for the importance of low-energy electrons in triggering tribochemical reactions, the most recent Nakayama and Nevshupa discovery of microplasma generation in a gap around a sliding contact should also be taken into account. Specific issue of this discovery  is that the plasma emits mostly invisible ultraviolet photons and, to a lesser extent, infrared photons. From the view-point of the microplasma components and the general triboemission and/or fractoemission processes it is also necessary to consider an interaction of negatively charged particles with photons. Bearing in mind all these facts Kajdas [8, 11] proposed to understand the flash temperature effect as the factor initiating tribochemical reactions because the interaction result of UV photons with negatively charged particles (clusters) leads to generation of additional low-energy electrons that, according to the NIRAM approach, are responsible for trigging tribochemical reactions. Thus, it is possible to conclude that low-energy electrons might be considered as the governing factor of tribochemical reaction initiation processes. Accordingly, flash temperatures might also be expressed in the form of electronic energy. This topic was discussed in detail in another paper of the conference book .
TRIBOCHEMISTRYThe mechanical action at solid surfaces tends to promote chemical reactions and produce surface chemistry that may be entirely different to those observed in static conditions. Chemical reactions are characterized by chemical kinetics concerning with their velocity and also the mechanism by which chemical reactions occurs. Their mechanism is applied to demonstrate the step-by-step sequence of the events that are postulated to proceed at the molecular level as reactants are changed to given organic, organometallic, or inorganic products. The clear mechanism should also consider all intermediates and/or transition states. The summation of all the changes that occur, expressed as net reaction, is not the whole situation because the net reaction change usually consists of several consecutive reactions. This is of particular importance for better understanding tribochemical processes. Chemical reactions of tribological additives proceeding during the boundary lubrication process involve the formation of a film on the contact surface which protects it during friction. These reactions initiated by mechanical action in the direct contact zone relate to tribochemistry dealing with the chemical changes of both solid mating elements and lubricant molecules due to the influence of boundary and/or mixed lubrication operating conditions. The chemical reactions under friction conditions include several types of reaction processes (i) genuine tribochemical reactions, i.e. specific chemical reaction process initiated only by the mechanical action of the system (mechanochemistry), (ii) reactions of thermochemical processes (e.g., oxidation, polymerization, degradation), (iii) heat enhanced tribochemical reactions relating to typical tribochemical process that is controlled by both mechanical action and heat, and (iv) catalytic processes related to both catalysis and tribocatalysis. Typical physical phenomena evolving from wearing processes comprise increased contact temperature and triboemission. Emitted low-energy electrons can initiate tribochemical reactions which govern tribology processes by forming wear reducing or corrosive wear films.
A relevant application of Kajdas` hypothesis was the development of the mechanism of vinyl-type tribopolymerization [5, 17-22], the seminal idea for which was earlier introduced by Furey . Another example considers reactive intermediates generated by low-energy electrons and UV irradiation. Zhao et al.  hypothesized that low-energy electron mediated degradation of PFPE lubricants under sliding conditions of computer head on hard disk in UHVC is the dominant mechanism. The hypothesis was very helpful in better understanding and accounting for all the major experimental findings. Work by Vurens et al.  demonstrated the negative-ion emission from the Demnum S200* lubricant during irradiation with low-energy electrons (e.g., of less than 14 eV). Figure 1 shows spectrum of negative ions produced from Demnum S200 lubricant by low-energy electrons. They found that the most intense fragment ion relates to F- at m/e = 19. This is not unexpected since F- is a very stable ion (hard base). Other major ions include C3F5O2- (m/e = 163) and C3F7O- (m/e = 185).
*Commercially available perfluoropolyether (PFPE) class of lubricants are known by brand names as Fomblin, Demnum and Krytox. Demnum lubricants have linear structures. They are liquid over a wide temperature range and completely inert to most chemical agents, including oxygen. The most sophisticated applications of these lubricants include magnetic recording disks for computer data storage.
The negative ion (m/e)- = 185 is generated via C-O bonding cleavage in one end of the Demnum S200 molecule. The mechanism by which ion (m/e)- = 163 is formed was proposed by Kajdas ; that work aimed at accounting for this specific ion formation mechanism. According to work  the formation mechanism of the negative species (m/e)- = 163 proceeds as follows:
Interestingly, the NIRAM approach can also be applied in accounting for experimental results concerning chemical bonding of PFPE lubricant films with DLC under sliding conditions . Review paper  discusses and analyzes the literature concerning the interaction and degradation mechanisms of PFRE lubricants with carbon protective overcoats used for magnetic media. Figure 3.
Another paper  related to some wear issues and a general degradation mechanism of fluorinated compounds. Most recently, Przedlacki et al.  described both tribochemical performance and lubrication mechanism of selected fluorinated alcohols containing 5 to 10 carbon atoms. The experiments were performed on an Optimol SRV tester with steel/steel and steel/Al contacts under boundary lubrication conditions. The products of the tribochemical reactions were investigated using FTIRMA, XPS and SEM/EDS techniques. An application of the NIRAM approach was proposed for the reaction path and structure of an oligomeric film formed by fluorinated alcohols under friction conditions. It was demonstrated that alcohols form anions or anion-radicals which in turn react with positively charged metal surface spots. In the case of the investigated fluorinated alcohols the process may proceed in two steps. The first step is the interaction between lubricant molecules and low-energy-electrons. The second is the reaction of anions and anion-radicals with positively charged surface followed by cross-linking between unsaturated anions and anion-radicals reacting with the surface, leading to an oligomeric organo-metallic layer which protects the surface.
Tribochemistry of alcohols, fatty acids, esters, ethers and hydrocarbons was recently discussed in . Majzner et al. , investigating tribochemical reactions of carboxylic acids under boundary friction conditions found that, apart from monodentate carboxylate group, salts with double bond in a, b position and chelating symmetric bidentate carboxylate group are formed. These results do not exclude the possibility that under boundary friction conditions carboxylic acids react on steel surface according to negative-ion-radical action mechanism. Figure 4 shows the proposed structure of iron salt with double bonding. Such compounds were not detected after reactions taking place under static conditions. This also clearly shows the difference between tribochemical and thermochemical reactions.
BRIEF DESCRIPTION OF THE NIRAM APPROACH
Due to action of the emitted electrons with the lubricant molecules, negative ions and radicals are produced on the rubbing surfaces. Interaction of negative ions with a solid surface leads to the formation of a chemisorbed protective film (organometallic film) which protects the rubbing surfaces from wear. If shear strength is high, it can cause cleavage of chemical bonds producing an inorganic film, for example iron sulfide, and generating further radicals followed by adequate chemical reactions.
Under severe operating conditions destruction of the inorganic protective film takes place causing generation new activated sites as demonstrated in Figure 5. Eventual destruction of protective layer by wear, followed by electron emission and subsequent formation of a new protective film according to stages (a) through (c) as depicted in Figure 7.
TRIBOEMISSION APPARATUSA new high-vacuum triboemission instrument - that is based on a Channel-Electron Multiplier (CEM) in the pulse-counting mode - was designed, constructed and employed for the following presented measurements of charge intensity. The new triboemission instrument features a rotating-sliding shaft which introduces the mechanical inputs into a high-vacuum chamber. A wide range of loads can be applied to the shaft by a weight and fulcrum. Also, a variable-speed electric motor can produce a wide range of rotational speeds. A biased-grid is placed between the emitting surfaces and the detector. A schematic diagram of the triboemission device is shown in Figure 8. Detailed descriptions of the triboemission instrument's features, operating ranges and measurement and data acquisition capabilities were presented in the work of Molina et al [33-35]. Figure 8a shows a photograph of the triboemission set-up.
REVIEW OF SOME EXPERIMENTAL RESULTSMain research work by the authors allowed to find significant emission of electrons from diamond-on-alumina, diamond-on-sapphire and alumina-on-alumina during sliding contact and after contact ceased, while no emission was detected from diamond-on-aluminum systems. Most recent research  was focused on negative-charge intensity from the semiconductors silicon and germanium while scratched by a diamond pin in high vacuum. In these experiments significant triboemission was always detected during the sliding contact, while emission was undetectable when contact ended. Emission intensity was substantially higher for diamond-on-Si than for diamond-on-Ge. Decreasing emission was observed when the same wear track was scratched in repeated passes.
The CEM-counting in the 10 msec acquisition-window (each count representing a single particle of enough energy to reach the detector) shows burst-type emission significantly higher than background-noise. Decaying emission, though of lower level, is also observed after the contact ceased.
Measurements of triboemission also were performed from alumina-ball on alumina-disk systems. A typical measurement of triboemission from alumina-on-alumina during contact and the post-contact emission is shown in Figure 10. Applied load was 2 N and the employed sliding speed was 0.14 cm/second during a 45second contact-period, after which contact ceased. The gaps of data acquisition in Figure 11 correspond to periods where data acquisition was set off to save acquired data at the employed acquisition rate of 10msecond-window.
Some small bursts of triboemission show for about 45 seconds after the starting of sliding contact, to be followed after this relatively long period of contact by the largest observed burst (of 603 counts in the 10 mseconds-window). This particularly large burst also initiated sustained high-level post-contact emission after the contact ceased. This research work provided the first known evidence that alumina-on-alumina systems emit charges in conditions of sliding contact, and that scratching by a diamond is not needed for triboemission to occur . This finding is of particular importance because it provides a clear evidence for the hypothesis concerning the action mechanism of vinyl monomers when used as antiwear lubricant additives to lubricate an alumina-on-alumina system . Interestingly, that an initial period of low negatively-charged emission may exist for alumina system, at least for the applied load and sliding speed; however, large bursts of high-intensity triboemission will follow if sliding contact is run on the same wear track; this is illustrated in Figure 11.
A study on the statistical significance of these results and others previously reported vs. CEM background can be found in the work of Molina . Such study demonstrated statistical significance for the measured average negatively-charged triboemission from diamond-on-alumina, diamond-on-sapphire and alumina-on-alumina versus the CEM background noise, while found no significant difference for the average positive-charge CEM counts from same material systems, or for the weak CEM counting from the diamond-on-aluminum contacts.
More detailed review on tribochemistry and development of the tribopolymerization concept along with experimental results on triboemission from ceramic sliding contacts is presented in work .
The most recent investigation by Molina et al.  relates to triboemission from the sliding contact of Si and Ge. Fracture of Si and Ge occurs by brittle cleavage after microcracking in a few crystallographic planes , and surface reconstructions for Si and Ge surfaces reduce to a few well-known stable structures [40, 41]. In addition to possessing simpler structures than those of ceramics, semiconductors present an advantage for triboemission studies because their surface-charging is negligible in comparison with charge-effects in insulators. Dickinson et al.  performed research on particle emission and transient electrical measurements of diamond/MgO contact. They investigated electron and photon emission from reciprocating scratching of MgO with diamond. Measurements included (i) negative-charge intensity and (ii) the rate of negative-charge triboemission vs. kinetic-energy by a channel electron multiplier (CEM) in vacuum. Interestingly, the rate of emission rapidly dropped for increasing energy and a large fraction of they total negative-charge emission was for energy lower than 100 eV. Fractoemission from semiconductors was also investigated by Dickinson et al. , who used a channel electron multiplier (CEM) in pulse-counting mode to detect electron-emission outputs from bending fracture of single-crystal Si. They also measured atomic and molecular Ge emission from Ge fracture . Kaalund and Haneman  studied in-vacuum emission during cleavage by bending of Si and Ge. They observed burst-type electron emission starting at the onset of cleavage with durations ranging from tens of microseconds to 1.8 milliseconds. The signals were independent of dopant concentrations, of high-vacuum levels and of temperature. Maxima of integrated signal intensity were up to three times higher for Si cleavage than for Ge under the same conditions. All the reported fractoemission outputs were substantially longer than the typical fracture durations. A review of relevant research work on triboemission can be found in the paper of Kajdas et al. .
Table 1 summarizes most relevant triboemission data obtained for the investigated material systems and their statistical analysis. It also presents for comparison some relevant triboemission data from alumina system obtained by Molina et al. [33-35, 38].
(a) Average rates for two different wear-tracks on same specimen in 120 sec-contact (four turns). Load: 5 N. Speed:1-1.9 mm/s. (b) Average rates for different specimens in 320 sec-contact. Load: 2-5N. Speed: 1.4-4.8 mm/s. (c) For differences between average triboemission count and CEM background. Sample size: 2 (3 for alumina-on-alumina). Molina et al.  presented the first-known evidence that negative-charges are emitted from diamond-on-Si and diamond-on-Ge during sliding contact. This burst-type electron emission is not substantially different from those shown in diamond-on-insulators sliding. A significant difference was found between triboemission rates from diamond-on-Si and diamond-on-Ge, and the possibility is discussed that such difference relates to the material hardness. Decreasing triboemission is also observed from the semiconductors when the same wear track is scratched in repeated passes, and such decreasing triboemission may relate to decreasing semiconductor wear. Clear evidence for a new finding is presented that no emission is produced after the contact ceased for these semiconductors, while post-contact triboemission was previously detected for an insulator (e.g., alumina) under vacuum.
NIRAM - HSAB Concept of Ceramic TribochemistryThe concept is based upon the ionization mechanism of ceramic lubricant compound molecules caused by the action of exoelectrons. As already mentioned before, the principal thesis of the model is that lubricant components form anions which are then chemisorbed on the positively charged areas of rubbing surfaces. Paper  presents a generalized NIRAM - Hard and Soft Acids and Bases (HSAB) theory and provides its possible application for accounting for some tribochemical processes under boundary conditions. The approach combines reactivity of lubricant components and tribological solids according to reactions induced by low-energy electrons and reactions affected by acid-base interaction. Major HSAB interactions are related to tribochemical reactions, initiated by low energy electrons. Figure 14 depicts a general approach to the NIRAM - HSAB action mechanism.
The difference can be explained in terms of the NIRAM-HSAB approach as depicted in Figure 16. This mechanisms distinctly demonstrates that the consumed molecule of water does not produce any gas. Of course further steps of the reaction chain will produce ammonia. Assuming that not all hydrogen radicals recombine with only one nitrogen atom, it is easy to realize that even over ten water molecules might be adsorbed to produce only one ammonia molecule. Further reaction steps relate to generation of -NH2 species and finally ammonia (NH3) is produced.
Basing on data presented in Figure 13, now it is possible to suggest that the tribochemical reaction of silicon with water is also governed by the NIRAM-HSAB approach. Accordingly, it is feasible to conclude that tribochemistry of ceramics mostly relates to chemical reactions initiated by triboelectrons in microscopic solid/solid contacts.
The authors plan to further develop their research in both the area of triboemission measurement and tribochemistry applications. The latter would particularly relate to tribochemistry of ceramics. It is their intention to explore several existing unknowns in the area of triboemission, and to extend their work to other materials of tribological application and some metal catalysts.