THE ROLE OF TRIBOEMISSION IN CERAMIC TRIBOLOGY: A REVIEWG.J. MOLINA, Assistant Professor
School of Technology, Georgia Southern University
Statesboro, Georgia 30460-8047, USA
gmolina @ gasou.edu
M.J. FUREY, Professor
A.L. RITTER, Professor
C. KAJDAS, Professor
AbstractA review is given of triboemission of charged particles from ceramics, including metallic oxides MgO and Al2O3, with the goal of providing insight into the role of triboemission in ceramic tribology. Experimental results are discussed for the initiation of triboemission in these materials by tensile deformation, bending, and abrasion. Possible models for these triboemission processes are discussed.
Keywords: triboemission, tribochemistry, fractoemission.
INTRODUCTIONMechanical energy associated with the boundary lubrication process produces a large number of material and energy outputs: such as, material debris, heat, acoustic emission, charged particles and various radiation. In particular, the emission of electrons, ions, neutral particles, photons, radiation and acoustic emission under conditions of tribological damage is called triboemission. Figure 1 shows a conceptual view of triboemission. The phenomenon of charged-particle emission from solid surfaces when subjected to mechanical work, wear, tensile deformation or surface reaction is called exoemission . Fractoemission arises from the tensile deformation of brittle solids [2,3].
Low-energy triboemitted electrons are thought to play a significant role in tribochemical reactions under wear conditions, particularly in the boundary-lubrication mechanism of tribopolymerization . The role of triboemitted low-energy electrons in tribochemical reactions was formally described by Kajdas et al. [5-7]. He proposed a concept of ionization of lubricant components by low-energy exoelectrons (1 to 4 eV), that he called the Negative-Ion Radical Action Mechanism (NIRAM). A key element of such mechanism is the emission of low-energy electrons from freshly formed surfaces in the friction of mating surfaces. The work of Molina et al.  reviewed the concepts of tribopolymerization and NIRAM.
Our paper is organized as follows. First, we discuss the measurement of triboemission with an initial section on instrumentation used by the various groups. Then a short review is given of relevant work by research groups at 1) the Mechanical engineering Laboratory, Tsukuba, Ibaraki, Japan, 2) Washington State University and 3) Virginia Tech. The section on experimental measurements closes with a summary of the experimental results. Then a discussion follows of various models for the triboemission process. These models are broken down into two groups: triboemission models for tensile deformation and triboemission models for abrasion of a surface. Three models are considered for the tribo-emission process arising from tensile deformation. They are the electrified fissure mechanism, strain energy at a crack tip and recombination on fractal networks. Finally, a more qualitative description of triboemission models for surface abrasion is given.
MEASUREMENTSSeveral techniques have been developed to study triboemission. These different instruments are described next. Then experimental results from various laboratories are reviewed.
InstrumentationTriboemission of charged particles is generally detected by channel electron multipliers (CEM). These detectors require an ambient pressure of less than or equal to 10 mPa. Photons are detected with a sealed photomultiplier tube so that the mechanical abrasion of the sample surface can be studied in atmospheric pressure or under flowing gases.
The research group at the Mechanical Engineering Laboratory, Tsukuba, Ibaraki, Japan pioneered a technique consisting of a metal ring current-collector mounted 0.5 millimeters above the tip of the scribing tip stylus [9,10]. The current-collector was biased +15V or -15V to detect negative or positive charge, respectively. The detected current could be either triboemitted charges passing from the abraded surface through the ambient atmosphere to the metal ring or surface current from the tip across the stylus surface to the metal ring. The observed currents ranged from 100 to 1000 pC/s which is three to four orders of magnitude higher than the typical current observed with a CEM (@ 105 charges/second or @ 0.02 pC). This suggests that the surface current is appreciable. The metal ring detector is an excellent way to study the creation of free charge by abrasion of the surface because it can be used at any pressure and in any gas. But since the relation between the observed current and triboemission is unclear, we have not included work involving this detector  in our review.
Research at Washington State University observed electron and photon emission generated by abrasion [12,13], three-point material fracture [19,24], and tensile deformation of the sample [19,22,23]. In their apparatus, electron and photon emission could be detected simultaneously with a CEM and a photomultiplier tube and the retarded-energy spectrum of the emitted electrons were measured with a retarding grid analyzer . They also did seminal experiments in which particle emission and the acoustic emission from the creation of cracks in tensile deformed oxide films were observed in coincidence [2,3].
Molina et al. developed a high-vacuum triboemission instrument to measure charge intensity and the retarded-energy spectrum of negatively- and positively-charged triboparticles. They used a CEM detector in the pulse-counting mode with standard pulse detection instrumentation and fast-rate data acquisition by computer. The instrument is described in Mo Una's thesis .
Mechanical Engineering Laboratory, Tsukuba. Ibaraki. Japan.The research of Nakayama and colleagues in Japan represents the most extensive and varied work in the field of triboemission. It is clear that many significant contributions were made by this group as already mentioned and as shown by the Nakayama references quoted in this paper. But for reasons mentioned above, our discussion here is limited to experiments involving CEM measurements with ceramic materials. Nakayama et al measured triboemission from an alumina ball sliding on amorphous hydrogenated carbon films in dry air at 5 MPa  with a CEM. The very low loads and relatively high speeds used are typical of magnetic-media lubrication technology. This work made clear the burst-type nature of charged-particle triboemission from insulators and the importance of wear behavior, not the coefficient of friction, in characterizing the observed phenomena. They also showed the influence of surrounding atmospheres on the triboemission intensity.
Washington State University. Pullman, Washington.This group found significant burst-type electron emission and photon emission (e.g., of respectively 3 x 106 and 2 x 104 counts/sec for largest peaks) [12,13] for abrasion of MgO. They observed differences between the highest spike level and the average emission as high as three orders of magnitude for photon and electron emission. They also found that electron emission from tensile deformed oxide films coincided with the creation and propagation of cracks [2,3]. The experimental conditions they used are summarized in Table 1.
Significant difference was found between the decay time for photon and electron emission bursts associated with abrasion of MgO. The photon emission bursts for diamond-on-MgO decayed on the ms-scale, returning to near-zero between spikes. The electron emission, however, showed a characteristic decay-time of about 100 ms after each spike and with overlap between emission from different spikes creating a quasi-continuous envelope. In simultaneous measurements of photon and electron emission under stick-slip conditions, they observed a 5 to 10 msecond-lag for the rise of electron emission with respect to that for photon emission. Evidence was reported of emission after the contact ceased, although of longer decay period for electron emission than that for photon emission.
Further evidence for the close correlation between electron emission and fracture of oxide materials is provided by an experiment in which MgO was indented by a diamond tip . Photon emission was observed from the onset of contact and deformation, while electrons were emitted only after surface fracture occurred. Dickinson et al.  also investigated the kinetic-energy of triboelectrons from diamond-on-MgO using the retarding-grid method. The rate of electron emission (in counts/sec) decayed monotonically with decreasing retarding-grid voltage dropping by over 90% from zero to -100V. Grid voltages up to -1,000V were not enough to completely stop the emitted electrons.
The same work by Dickinson et al.  found that both diamond and MgO produced the electrons, although the net surface-charge (measured by an electrometer) was negative and positive for diamond and MgO, respectively. Since there is no known mechanism for electron emission from diamond, it was suggested that emission came from MgO debris that had transferred to the diamond tip. The authors admit that it is a mystery how electron emission could take place from a wear track with net positive surface.
Virginia Tech. Blacksburg. Virginia.The group at Virginia Tech measured emission from diamond-on-alumina, diamond-on-sapphire and diamond-on-aluminum, and from alumina-on-alumina [14,15]. Thev used two different contact geometries: a cone diamond-pin sliding on rotating-disks of three related materials (e.g., alumina, sapphire and aluminum), and an alumina-ball sliding on alumina-disk. The experimental conditions used are summarized in Table 1.
For the three ceramic-systems (e.g., diamond-on-alumina,diamond-on-sapphire and alumina-on-alumina) they observed burst-type negative-charge triboemission at a load of 2N and a surface speed of 0.48 cm/s. They found no significant difference between emissions from polycrystalline alumina and sapphire. The emission rate decayed monotonically after the contact ceased for all three ceramic systems and the largest bursts of negative charge occurred during the first two minutes of contact. Smaller bursts of emission, lasting several minutes, were a regular feature of the emission. No negative emission above the background was observed for diamond-on-aluminum. Nor did they observe any positive emission from any of the systems studied.
The retarded-energy spectrum of negative charge emission from diamond-on-alumina and diamond-on-sapphire was measured and it was found that a significant fraction of the emission was in the range 0-5 eV with detectable emission extending to the minimum retarding voltage of —48 volts.
The work of Molina et al. is relevant to the triboemission field in (a) obtaining the first known evidences that alumina-on-alumina emits triboelectrons; it is not necessary to scratch a solid with a diamond to produce emission, (b) demonstrating that the measured negatively-charged triboemission is composed of low-energy electrons and (c) relating triboemission measurements to tribochemical reactions. Their experimental results are consistent with the NIRAM hypothesis of Kajdas (that low-energy electrons, and no positive ions, are significantly trihoemitted from alumina).
SummaryTable 1 summarizes the material and detection systems used, atmospheres, applied loads and sliding speeds, and maximum emission intensities measured for selected research on charged-particle triboemission.
A common feature of these measurements is the random nature of the emission as a function of time with intervals of intense emission that is two or three orders of magnitude greater than the average emission. The group at Washington State University has clearly demonstrated that a primary source of emission from oxide materials is the creation and propagation of cracks. The line of evidence is as follows. They have observed coincidence between acoustic emission from cracks and particle emission when the sample is strained. When a sample is indented, the emission of photons is high and the emission of electrons is very low while the sample is plastically deformed, but then the emission of electrons increases by an order of magnitude when fracture occurs. Finally, when a sample is abraded under stick-slip conditions, the emission of photons, but not electrons, is observed when the stylus first begins to slip by plastic deformation. Then electron emission turns on when the stylus starts to plow through the material by fracture. And finally, emission ends when the stylus stops slipping. Though formation and propagation of cracks in oxide materials is a necessary condition for triboemission, it may not be a sufficient condition.
MODELS OF THE TRIBOEMISSION PROCESSThere are three modes of surface deformation that may give rise to triboemission: tensile deformation, bending, and abrasion. Tensile strain of an oxide coated metal gives rise to emission of particles and photons. Fracture of single crystal MgO and MgF2 by bending also results in the emission of particles. A clear correlation exists between the creation of cracks during tensile strain and the emission of particles. Less clear is a microscopic mechanism for the emission of particles from cracks and several models have been suggested. These models are reviewed below. If an oxide surface is abraded by a hard tip, generally diamond, triboemission is also observed. Abrasion is considerably more difficult to characterize than tensile deformation or fracture and no microscopic models exist for emission due to this process. It is highly likely that the creation of cracks due to abrasion gives rise to triboemission, but other mechanisms may contribute also. A brief discussion of possible models for emission from abraded surfaces is given in section 3.2.
Triboemission models for tensile deformationWe identified three microscopic models that suggest a mechanism for creating triboemission from cracks that are produced by tensile deformation of oxide films or bending of single-crystal oxide films. We label these models: (i) The electrified fissure mechanism, (ii) the localized strain energy mechanism, and (iii) the recombination mechanism. In addition, we find two models that attempt to describe the statistical nature of the emission process. Since the connection between these statistical models and tribological applications is tenuous, we have not considered them in this review.
Electrified fissure mechanismThe correlation between crack formation and triboemission and the absence of emission from pristine, oxide-free metal surfaces suggest that the emission may be due to surface charge created on the walls of rapidly propagating cracks. The breaking of ionic bonds in insulators by crack formation will leave opposite charge on the two sides of the opening crack. The initial expression of this electrified fissure model suggested that triboemission was a consequence of the electric field created by the separated sheets of charge, but did not provide details regarding an emission mechanism [16-18]. Dickinson et al. expanded this fissure model taking into account the interaction between the electric field created in the crack and residual gases on the crack surface .
Strain energy at a crack tipAmott and Ramsey  noted that the strain energy at the tip of a propagating crack could provide the necessary energy to eject particles from the surface of a solid, but they did not suggest a specific microscopic mechanism for coupling the strain energy to an individual ejected particle. Rosenblum et al.  considered this model for triboemission in more detail and found that the excess strain energy at the tip, about 1/4 eV per atom, is the right order of magnitude for ejecting particles. But a microscopic mechanism for coupling this strain energy to the single ejected particle is still lacking. Rosenblum et al. found that the peak emission intensity as a function of strain did depend strongly on the thickness of the metal oxide layer and successfully explained this dependence in terms of their crack tip model of triboemission.
Recombination on fractal networksA consequence of the electrified fissure mechanism for electron emission is the creation of electron-hole pairs on crack surfaces. Dickinson et al.  find that the decay kinetics after fracture depend strongly on the local geometry of the fracture surface. The electrons diffuse by thermally stimulated short-range hopping until trapped by defect states near the conduction band edge. The trapped electron recombines with a hole in the valence band and may emit an electron by an auger-like process or a photon by radiative recombination. These two processes for electron and photon emission can be modeled by recombination processes of the form: A + B ® C + photon/electron where A is the mobile charge and B is a trap.
Triboemission models for abrasion of a surfaceThe generation of triboemission by scratching a surface is considerably more complex than triboemission generated by bending or tensile strain of a material. Processes that might contribute to triboemission from a single pass of the scribe over a surface include: plastic deformation, fracture, surface excitations, and chemical reactions between the atmosphere and freshly exposed surfaces. Additional processes that might contribute to triboemission from multiple passes of the scribe over the same wear track include: wear transitions to different modes of fracture and material removal, crushing and grinding of wear debris, and thermal load. A preliminary study of Al2O3 abraded with a diamond tip does observe a transition in the electron emission rate after multiple passes. A hypothesis is being tested that the emission transition is correlated with wear transitions in this ceramic [25,26]