...

Hsu-Gates2001-BoundaryLubrications.pdf

by user

on
Category: Documents
5

views

Report

Comments

Transcript

Hsu-Gates2001-BoundaryLubrications.pdf
12
Boundary Lubrication
and Boundary
Lubricating Films
12.1
Introduction
Definitions
12.2
The Nature of Surfaces
Surface Structures and Compositions • Surface Energy and
Reactivity • Surface Emission under Stress • Surface
Roughness and Relative Conformity
12.3
Lubricants and Their Reactions
Lubricant Basestocks and Additives • Relationship Between
Oxidation Reactions and Film Formation • Organometallic
Chemistry and Tribochemistry
12.4
Physical and Chemical Properties • The Detection of
Organometallic Compounds in Films • Mechanical
Properties of Boundary Lubricating Films • Advances in
Measurement Techniques
Stephen M. Hsu
National Institute of Standards
and Technology
12.5
Richard S. Gates
National Institute of Standards
and Technology
12.1
Boundary Lubricating Films
Boundary Lubrication Modeling
Wear • Flash Temperatures • Asperity-Asperity
Understanding • Molecular Dynamics Modeling
12.6
Concluding Remarks
Introduction
Lubrication may be defined as any means capable of controlling friction and wear of interacting surfaces
in relative motion under load. Gases, liquids, and solids have been used successfully as lubricants.
Boundary lubrication usually occurs under high load and low speed conditions in bearings, gears, cam
and tappet interfaces, piston rings and liner interfaces, pumps, transmissions, etc. In many cases, it is
the critical lubrication regime that governs the life of the components subject to wear. Because of its
industrial significance, many studies have been conducted in the past. The most comprehensive was the
1969 assessment by the American Society of Mechanical Engineers (ASME) Research Committee on
Lubrication (Ling et al., 1969). The study included critical reviews on surface physics, chemistry, fluid
mechanics, contact mechanics, and materials science. The major conclusion in that review was that more
research was needed to understand the complex chemical, physical, and material interactions. Since then,
topical symposia have been organized by different researchers on lubricant chemistry, contact mechanics,
microelastohydrodynamic lubrication (µ-EHL), analytical techniques, boundary films, and molecular
dynamics simulations. Cross communication and integration of the significant advances in analytical
chemistry, surface analysis, materials sciences, and molecular modeling, however, have seldom been made.
This chapter attempts to provide a bird’s eye view across these disciplines.
© 2001 by CRC Press LLC
12.1.1 Definitions
When the contact geometry and the operating conditions are such that the load is fully supported by a
fluid film, the surfaces are completely separated. This is generally referred to as the hydrodynamic
lubrication. Theory for fluid film design is well developed based on Reynolds’ equations and continuum
mechanics (Reynolds, 1886).
When the load is high and/or the speed is low, the hydrodynamic or hydrostatic pressure may not be
sufficient to fully support the load, and the surfaces come into contact. The contact occurs at the peaks
and hills of the surfaces, and these are referred to as asperities. The amount and the extent of the asperity
contact depend on many factors: surface roughness, fluid film pressure, normal load, hardness, and
elasticity of the asperities, etc. Many of the asperities undergo elastic deformation under the contacting
conditions, and the normal load is supported by the asperities and the thin fluid film. This condition is
generally referred to as the elastohydrodynamic lubrication (EHL) (Dowson and Higginson, 1959). The
EHL theories are reasonably well developed and are capable of describing the surface temperatures, fluid
film thickness, and the fluid film pressures supporting the load. The theory assumes continuum mechanics and does not take into account the effects of wear and the presence of a third body (wear debris). No
chemical effect between the asperity and the lubricant is taken into account.
Further increase in the contact pressure beyond the EHL conditions causes the contacting asperities
to deform plastically and the number of contacts to increase as well as for the fluid film thickness to
decrease. When the average fluid film thickness falls below the average relative surface roughness, surface
contact becomes a major part of the load supporting system. Mechanical interactions of these contacts
produce wear, deformation, abrasion, adhesion, and fatigue under dry sliding conditions. Chemical
reactions between the lubricant molecules and the asperity surface, due to frictional heating, often
produce a boundary chemical film which can be either beneficial or detrimental in terms of wear. The
combination of the load sharing by the asperities and the occurrence of chemical reactions constitutes
the lubrication regime commonly referred to as the boundary lubrication (BL) regime. Figure 12.1 shows
an approximation of the relationship for these regimes as they relate to coefficient of friction and contact
severity.
Under boundary lubrication conditions, interactions between the two surfaces take place in the form
of asperities colliding with each other. These collisions produce a wide range of consequences at the
asperity level, from elastic deformation to plastic deformation to fracture. These collisions produce
friction, heat, and sometimes wear. Chemical reactions between lubricant molecules and surfaces usually
accompany such collisions producing organic and inorganic surface films. It has long been thought that
surface films protect against wear. Closer examination (Hsu, 1991) suggests that some films are protective
(antiwear), some films are benign, and some films are detrimental (prowear).
A theory for a comprehensive view on boundary lubrication is currently lacking; however, models on
lubricant chemistry and contact mechanics do exist (Klaus et al., 1991; Blencoe et al., 1998; Yang et al.,
1996; Cheng and Lee, 1989). Recently, molecular dynamics models have been developed to describe
atomic interactions under simplified boundary lubricated conditions (Stuart and Harrison, 1999). The
detailed physical and chemical processes occurring in the contact zone are still not well understood.
Where does EHL end and BL begin? Since the surfaces have a range of asperity height distributions,
two surfaces in contact produce a range of distribution of stresses within the contact zone. Therefore, in
practical systems there is often no pure EHL or BL lubrication regime, and a mixed lubrication regime
exists. Some asperities are in the hydrodynamic mode, some asperities in EHL, and some asperities in
BL mode. As wear occurs, surface topography also changes. Depending on the nature and extent of
chemical reactions, conformity of surfaces can either develop or disappear. This changes the real area of
contact and, hence, the asperity stress distribution. The following sections will attempt to present the
current view of boundary lubrication and address some of the important issues. We will discuss the
following topics: the nature of surfaces, lubricants and their reactions, boundary lubricating film formation, and modeling and prediction of boundary lubrication.
© 2001 by CRC Press LLC
FIGURE 12.1
12.2
Comparison of lubrication regimes encountered under different contact severities.
The Nature of Surfaces
12.2.1 Surface Structures and Compositions
Engineering materials usually go through a series of manufacturing/fabrication and machining/polishing
steps to become load-bearing components in machinery. These steps invariably change the surface
structure and sometimes the chemical composition.
For most engineering surfaces, the surface is covered with oxides. In the case of iron-based alloys, for
example, the surface is covered with oxides of iron such as FeO, Fe3O4 , and Fe2O3. The subsurface beneath
the oxide is often a deformed or case-hardened layer, often called the Beilby layer, which is the result of the
machining and polishing steps and/or heat treatment the material has undergone during the manufacturing
process. The microstructure of this layer generally is a microcrystalline phase dispersed in an amorphous
phase for most steels. Specific compositions depend on the particular alloying elements present.
© 2001 by CRC Press LLC
Contaminants in the alloying elements often diffuse to the surface during the manufacturing process,
resulting in higher concentrations of the minor elements near the surface than the bulk material.
The atoms in solids are bonded together by different forces. They are: covalent, ionic, metallic,
hydrogen and van der Waals. Covalent bonds are strong bonds that are formed when electrons are shared
between atoms of similar electronegativity. They have an intrinsic directional nature. Ionic bonds result
when complete electron transfer takes place between atoms of different electronegativity. Coulombic
attraction between the resulting charged ions results in strong bonds. Metallic bonds are the result of
attractive coulombic forces between positive metallic atoms with commonly shared free negative electrons
resulting in high electrical and thermal conductivity. Hydrogen bonds refer to when hydrogen acts as a
bridge between its primary bond to an electronegative atom and another electronegative atom. van der
Waals forces are weaker long-range forces that result from dipole–dipole interactions between atoms;
they influence bonding strengths in some solids. These bonds have different strengths and require
different amounts of energy to break them. Additionally, in a solid, complex atomic and molecular units
must fit together with a periodicity that minimizes electrostatic repulsive forces. When different bonds
are broken (by mechanical or chemical means), the resulting surface energy and reactivity are different.
For single crystalline solids, different crystalline planes often exhibit different physical, mechanical,
and chemical properties. Frictional studies have shown that different crystalline phases produce different
frictional resistance to sliding when subjected to the same conditions (Buckley, D.H., 1981). Thus,
polycrystalline solids as well as polymeric solids often have anisotropic properties. This adds to the
complexity of the surface structure/properties relationships. Under boundary lubricated conditions,
different crystalline phases also exhibit different reactivity to lubricants, moisture, oxidation, and surface
contaminants under sliding conditions.
12.2.2 Surface Energy and Reactivity
Surface energy for most engineering surfaces is difficult to determine. Rabinowicz (Rabinowicz, 1965)
has suggested that the surface energy of a solid can be approximated by using the surface energy of the
liquid at the melting point of the material. The agreement between this approximation and experimentally
determined value for simple pure materials is surprisingly good. It was further shown that the ratio of
the surface energy to penetration hardness could be correlated to friction, i.e., lower the ratio, lower the
friction.
Thermodynamically, the higher the surface energy, the more reactive the surface will be. Therefore
knowing the surface energy will give an indication of the potential chemical reactivity toward oxygen,
water, and lubricants. Unfortunately, as has been discussed before, the surface is covered with oxides,
other impurities, and mechanically altered layers. When the surface roughness is superimposed on it, the
energy of the engineering surface is very difficult to estimate.
Additionally, as with any solids, there are defects. Atomic misalignment, lattice mismatch defects,
dislocations, voids, and phase segregations abound in solids. When these defects are on the surface,
including steps, twists, kinks for crystalline solids, they provide high-energy sites at which chemical
reactions, adsorption, and catalysis take place preferentially (Yates et al., 1996).
Experimentally, contact angle measurements with a series of well-known liquids have been used to
estimate “surface energy” of a smooth solid surface. Results are useful to understand wettability and
spreading of liquids on that solid surface, but caution must be used in generalizing to other surfaces.
12.2.3 Surface Emission under Stress
When atomic bonds are broken at the surface by mechanical processes, energy is both consumed and
released. This often changes the surface energy state. When all the oxides and other mechanically
deformed layers are broken through by scratches, nascent surface emerges. The energy state of such
nascent surfaces is high compared with the “normal” surfaces.
© 2001 by CRC Press LLC
Simoi et al. (1968) suggested that the rubbing of metal surfaces produced highly localized electron
clouds, referred to as the “exo-electrons.” The presence of these electrons had also been detected by others
(Ramsey, 1967; Rosenblum et al., 1977). How this electron cloud could induce reactions that otherwise
would not occur remained to be proven. Lenahan (1990) reported the measurement of dangling bonds
on silicon nitride, suggesting that rubbing might produce dangling bonds on crystalline surfaces, and
these dangling bonds were highly energetic and reactive. Nakayama (1991) detected the emissions of
charged particles, electrons, and photons from scratched silicon-based ceramics. Dickinson et al. (1993)
observed the emission of particles and electrons from fracture of crystalline surfaces.
In repeated single-pass experiments, Ying (1994) observed the surface to be highly strained and ordered
for various metals. Strain-induced fracture and deformation appeared to be the dominant mechanisms
under lubricated conditions for metals. Would this highly strained state upon fracture produce electrons
and charged particles?
12.2.4 Surface Roughness and Relative Conformity
Surfaces have microscopic roughnesses which are often random in nature. Under concentrated contact,
these asperities are deformed either elastically or plastically to form the interface. So the initial interface
depends on the relative hardness and the relative roughness of the two surfaces in contact. If the two
surfaces conform perfectly to each other and all the asperities are deformed 100%, then the real contact
area is equal to the apparent contact area. Of course, this is not the case for most engineering contacts.
For highly loaded steel bearing surfaces under sliding conditions, the real area of contact may be only
15 to 25% of the apparent area of contact (Wang et al., 1991) depending on other parameters, such as
relative surface roughness and the relative surface hardness. If the normal force acting on the surface is
very high, then some plastic deformation will also occur. Greenwood and Williamson and others have
studied this topic extensively (Greenwood and Williamson, 1966; McCool, 1986).
On each asperity, there are subasperities which are smaller in scale. On each subasperity, there are
sub-subasperities, and so on. If the distribution of asperity heights is Gaussian in nature, such as in the
case of ball bearings, the typical surface roughness parameters such as average roughness (Ra), root mean
square roughness (RMS), skewness, and kurtosis, will be adequate in describing the surface roughness.
However, if the asperity height distribution is not Gaussian, as for example, when two or three major
peaks have many smaller-scale subasperities, then the multimodal characteristics of the load-bearing
asperities become important.
In lubrication, the surface roughness is an important indicator as well as a significant parameter in
the calculation of oil film thickness. Wang et al. (1991) suggested that in a lubricated case, even though
the surfaces may be rough, if they conform to each other under plastic yielding, then the relative surface
roughness may be quite small. Oil film calculations based on elastohydrodynamics may yield significantly
smaller film thickness than is actually present. Conformity has been demonstrated to change with time,
materials, chemical additives, and wear modes (Wang et al., 1991).
12.3
Lubricants and Their Reactions
In lubricated systems, liquid lubricants are commonly used. When liquid lubricants are used in engines
and machineries, they serve multiple functions. They control friction and wear, cool the surfaces, remove
debris and contaminants, generate hydrodynamic pressures to support load, and redistribute stresses
over the surface.
Since most of the liquid lubricants are hydrocarbons, they tend to oxidize, thermally decompose, and
polymerize. These reactions produce high-molecular-weight reaction products which lead to the formation of “friction polymers.” To understand the interplay between these reactions and wear, we need to
understand the chemical reactions that take place as well as the chemical composition of the lubricants.
© 2001 by CRC Press LLC
12.3.1 Lubricant Basestocks and Additives
Modern liquid lubricants are compounded from lubricating basestocks and chemical additives. Selective
combination of the two produces a myriad of special lubricants designed for different applications. There
are two major classes of basestocks: (a) petroleum or mineral-oil derived; and (b) specially synthesized
basestocks.
12.3.1.1 Petroleum Basestocks
Petroleum basestocks are selected hydrocarbon fractions derived from crude oils. They generally consist
of molecules containing 18 to 40 carbon atoms in three basic hydrocarbon types: (a) paraffins,
(b) aromatics, and (c) naphthenes (cycloparaffins). Most of the molecules are of the mixed type containing two or more basic hydrocarbon structures. These basestocks also contain a small percentage of
compounds containing heteroatoms, such as sulfur, nitrogen, or oxygen, substituted into the various
hydrocarbon structures. Typical molecular structures are illustrated in Figure 12.2. Stereochemical possibilities provide an astronomical number of structural variations in such molecules. This is one of the
reasons why the effects of molecular structure on lubrication are not well understood today.
In typical basestocks, the mass fraction of aromatics usually ranges from 5 to 40% with the average
about 20%; straight chain paraffins usually range from 10 to 20%; and cycloparaffins make up the
difference. Molecules containing heteroatoms (N, S, O) usually range from 0.5 to 4% depending on crude
source, processing technology, and viscosity grade. Although the heteroatoms are a small fraction of the
basestock mixture, they have a significant influence on basestock stability and friction. The aromatics
provide solvency for additives and oxidized products, but they tend to react to form oil insoluble products.
FIGURE 12.2
Base oil molecular structures.
© 2001 by CRC Press LLC
12.3.1.2 Synthetic Basestocks
Synthetic basestocks are mostly long-chain molecules produced by chemical reactions in order to obtain
specific characteristics. They are made from petrochemicals, animal and vegetable oils, and coal-derived
feed stocks.
There are five major classes of synthetic basestocks: synthesized hydrocarbons, esters, ethers, halogenated compounds and silicone polymers. Others include sodium–potassium eutectics, and inorganic
polymers of boron, phosphorus, and nitrogen, for highly specialized applications. More comprehensive
descriptions of synthetic lubricants are available in the literature (Gunderson and Hart, 1962; Shubkin,
1993).
a. Synthesized Hydrocarbons
Synthesized hydrocarbons include alkylbenzenes, cycloaliphatics, poly-α-olefins, and polybutenes. Alkylbenzenes are mainly used in low-temperature applications as hydraulic oils, greases, and sometimes
engine oils. Poly-α-olefins have found increasing acceptance due to their similarity to and compatibility
with petroleum base oils, good thermal stability, and excellent viscosity–temperature relationship.
b. Esters
Esters are by far the most common synthetic basestock. Large quantities of materials with various
structures are readily available from chemical manufacturers. They consist of monoesters, dibasic acid
esters (adipates, azelates, dodecanedioates), polyol esters (neopentyl esters), polyesters, phosphate esters,
and silicate esters. Most of these exhibit high boiling points, excellent viscosity–volatility characteristics,
and high-temperature stability. Esters generally exhibit good solvency and have good friction and wear
characteristics. They tend to react to form acidic species (acid esters and half acid esters) faster than
paraffins.
c. Polyethers
Among polyethers, polyglycol ethers provide a variety of viscosity and molecular-weight grades. They
are used mostly in water-based fire retardant hydraulic fluids, brake fluids, and rubber molding lubricants.
They are available in water-soluble and oil-soluble forms. They have low pour points, good compatibility
with rubber, and good sludge and varnish resistance. The volatility and oxidation stability of the polyethers are generally similar to those of petroleum basestocks. They have a unique property in that their
decomposition products are low-molecular-weight compounds similar in physical properties to the
original starting material. Polyphenyl ethers represent a thermally stable class of aromatic compounds
containing no paraffinic side chains. These materials exhibit relatively poor viscosity–temperature and
low-temperature fluidity properties.
d. Halogenated Compounds
In halogenated compounds the hydrogen is replaced by chlorine, fluorine, or bromine. This generally
results in reduced flammability. Nonflammability may be achieved by incorporating more than 60% of
halogen into the molecule. Toxicity has placed severe limitations on the use of chlorinated and brominated
compounds.
Perfluoropolyethers are examples of halocarbons that are in current use as lubricants. These materials
show excellent oxidation stability and good viscosity and volatility properties. The very high cost of these
materials has limited their use in commercial applications.
e. Silicone Polymers
Silicone polymers were one of the earliest compound types investigated for lubricant applications. They
are mainly dimethyl silicone polymers and methylphenyl silicone polymers with various chain lengths.
Methyl silicones were developed in the 1940s as the liquids with the best physical properties (viscosity–temperature) for lubricant application. Since then a series of modified structures, including phenyl,
phenyl methyl, chlorophenyl methyl, fluoromethyl, and alkyl (C4-C5) silicones have been developed to
overcome the major problems of the silicones, which is the lack of boundary lubricity on ferrous bearing
© 2001 by CRC Press LLC
systems. Silicones have excellent fluid properties and have been used as lubricants in systems designed
to run under hydrodynamic and elastohydrodynamic conditions where the bearing systems are nonferrous. They have probably the best viscosity–temperature characteristics of any synthetics, with very low
volatility and excellent oxidative and thermal stability. However, their solvency is poor and they are not
miscible with petroleum oils.
12.3.1.3 Additives
Most lubricants contain special chemical additives to impart specific properties to the basestocks to
enhance performance and inhibit degradation. A wide variety of additive chemicals has been developed.
There are 10 major classes of additions, categorized according to their functions: antiwear additives;
friction modifiers; extreme-pressure additives; detergents; dispersants; antioxidants; corrosion inhibitors;
pour-point depressants; defoamants; and viscosity-index improvers. There are also other additives such
as tackiness agents, fatty oils, thickening agents, and color stabilizers. Smalheer and Smith (1967) describe
many chemical types in each class. For additional information, see Booser (1984). Recent advances in
the additive area are available through patent summaries published periodically (for example, Ranney,
1980; Satriana, 1982).
a. Antiwear Additives
Additives are used to reduce wear under elastohydrodynamic and boundary lubrication conditions. The
most widely used are zinc dithiophosphates and tricresyl phosphate. Each of these additives has a family
of derivatives modified for different temperature and stability requirements. Other additives used for wear
control are acid phosphates, phosphites, sulfurized terpenes, sulfurized sperm oils, metal dithiocarbamates,
and occasionally some sulfides. The antiwear additives generally function by adsorbing on or reacting with
metal surfaces to form a protective film that is easily shearable. Since direct reactions or interactions are
sometimes involved, additive–metal compatibility is important. One lubricant with an antiwear additive
may function very well for one material pair but may fail in another material combination.
b. Friction Modifiers
Additives are also used to modify the frictional characteristics of the material under boundary lubrication
contact. Fatty oils are sometimes used, such as lard oil, tallow, sperm whale oil, porpoise-jaw oil, and
blown rapeseed oil. Recent research on friction modifiers has centered on glycerides, oil-soluble molybdenum compounds, finely dispersed graphite in oil, and synthetic esters of various fatty acids.
c. Extreme-Pressure Additives
Extreme-pressure (EP) additives are used under highly loaded conditions to prevent seizure, scoring, and
welding. In metal forming and cutting applications, wear of one contacting surface is acceptable. The
additives therefore are chemically active compounds containing chlorine, phosphorus, and/or sulfur.
Examples are chlorinated wax, sulfurized fatty oils, sulfurized mineral oil, chlorinated mineral oil, phosphosulfurized fatty oils, benzyl and chlorobenzyl disulfides, and some phosphites.
d. Detergent/Dispersant Additives
Detergent/dispersant additives may be separated into two classes of compounds: metal-containing (ashcontaining) high-temperature detergents and ashless low-temperature dispersants. Their common function is to prevent deposits from forming on metal surfaces due to oil oxidation, contamination, or
polymerization. The detergents function as surfactants that tend to adsorb on surfaces forming micelles
around the insoluble material from the degradation of the oil. A stable microemulsion results from this
process. Modern automotive and diesel engine lubricants contain some “overbasing” associated with the
detergents. Calcium carbonate is added in the form of micelles to the detergent. This form of overbase
can thus neutralize acidic components from either the environment or oxidation of the lubricant. There
are many classes of detergents: metal sulfonates, metal phenates, metal salicylates, and metal phosphonates
and thiophosphonates. Most of these can be overbased to varying degrees. Metals commonly used in
detergents are calcium, magnesium, barium, and zinc.
© 2001 by CRC Press LLC
Dispersants are primarily additives used to disperse oxidized oil-insoluble products, water, fuel, or
other contaminants at relatively low temperatures (100°C or below). Because they do not contain metal,
they sometimes are referred to as ashless dispersants. There are two main families: Mannich reaction
products and succinimides. Mannich reaction products are reaction products from polybutene and
phenol; they are treated with amines and boric acid. Succinimides are usually n-substituted long-chain
alkenyl succinimides. The detailed mechanism of the dispersion is not well understood, but most researchers agree that they adsorb on the oil-insoluble submicrometer particles and keep them finely dispersed
in oil without further aggregation.
e. Antioxidants
There are numerous antioxidants available for lubricant use. The limiting factor sometimes is oil solubility. The important classes are hindered phenols, amines, and sulfur and phosphorus compounds.
Hindered phenols are phenols in which the hydroxyl group is sterically blocked or hindered, such as 2,6di-tert-butyl-4-methylphenol. They act as peroxide-radical traps to interrupt the oxidation chain reaction.
Amines such as N-phenyl α-naphthylamine are also widely used. Sulfur and phosphorus compounds are
usually used at high temperatures in the presence of metals, which often catalyze oxidation of hydrocarbons. Some researchers speculate that sulfur and phosphorus compounds inhibit oxidation by passivating
the metal surface with a protective film. Some phosphorus compounds, such as zinc dialkyldithiophosphate, also act as peroxide-radical traps to stop the oxidation chain reaction. Johnson (1975) lists over
100 antioxidants used in lubricants.
f. Corrosion Inhibitors
Corrosion inhibitors are additives that protect metal components used in engines and bearings against
attack from acidic contaminants in the lubricant. Rust inhibitors are a subset of corrosion inhibitors
designed specifically to protect ferrous materials against attack. Corrosion inhibitors function by forming
a tight passive film on the metal surfaces to withstand the detergent or dispersant often present in the
same lubricant. The most common corrosion inhibitors are zinc dithiophosphate, zinc dithiocarbamate,
sulfurized terpenes, and phosphosulfurized terpenes. Common rust inhibitors are alkenyl succinic acids,
alkyl thioacetic acids and their derivatives, imidazolines, amine phosphates, and acid phosphate esters.
In high-temperature applications such as the internal-combustion engine, calcium or magnesium sulfonates are usually used.
g. Viscosity-Index Improvers
Viscosity-index (VI) improvers are polymers with molecular weights on the order of 100,000 or more,
which thicken the oil at high temperatures. There are numerous variants of these additives, but the most
widely used include three types: polymethacrylates, olefin copolymers, and polyisobutenes. These VI
improvers are currently used in automotive crankcase oils, automatic transmission fluids, hypoid gear
oils, and hydraulic fluids. The polymers tend to break down permanently with use and the viscosity of
the polymer solution is decreased irreversibly at high shear rates. The stability requirements for the
various areas of application are met by careful control of the molecular weight and concentration of the
polymer used. In order to change the viscosity–temperature relationship sufficiently, relatively large
dosage is required (5 to 10% neat polymer).
12.3.1.4 Lubricant Formulation Technology
The large array of available petroleum basestocks, synthetics, and various chemical additives makes
numerous combinations possible. Lubricant formulators, when confronted with a particular lubrication
problem, select a certain basestock to satisfy the viscosity–temperature requirements and a particular set
of additives to give different characteristics. However, additives often interact with the basestocks as well
as with other additives in ways that are not understood. Therefore, most lubricants are developed by
actual equipment testing through trial and error. O’Connor (1968) lists over 30 major types of lubricants
and their requirements and some typical additives used in each application.
© 2001 by CRC Press LLC
12.3.2 Relationship Between Oxidation Reactions and Film Formation
The relationship between lubricant reactions and wear has long been observed in engines. Historically,
however, the link between the two technical areas is very weak. This is because the studies are largely
conducted by two separate communities. One group focuses on the oxidation mechanism and bench test
development; the other emphasizes the boundary lubrication mechanisms.
12.3.2.1 Chemical Reactions
The chemistry in the contact is complex and abundant. There are oxidation and thermal reactions of
hydrocarbons, polymerization to form high-molecular-weight products, adsorption and corrosion reactions, and catalysis by metal surfaces. In addition, tribochemistry, tribomechanical reactions, surface
oxide reactions, phase transitions, and mechanochemical effects also come into play. For semiconductors,
double-layer surface charges complicate the matter by introducing potential electrochemical reactions.
12.3.2.2 Oxidation and Thermal Reactions
It is well known that hydrocarbons oxidize via a free radical mechanism. The basic reaction mechanisms
are simple in principle but complex in reality. It follows primarily four steps: initiation, propagation,
branching, and termination (Hucknell, 1974):
RH → R• + H
1. Initiation
2. Propagation
R• + O2 → ROO•
ROO• + RH → ROOH + R•
ROOH → RO• + HO•
RO• + RH → ROH + R•
HO• + RH → H2O + R•
3. Branching
4. Termination
 alcohols
R• +R•
 aldehydes
R• +ROO•
⇒ 
ROO• +ROO•  ketones
 acids
RO• +R•
12.3.2.3 Metal Catalyzed Oxidation Reactions
Many metals, such as iron, chromium, copper, and nickel, have known catalytic activities with hydrocarbons. In lubricant oxidation studies, metal coupons are routinely used to simulate the catalytic
reactions encountered in actual applications. Studies reveal that the hydrocarbons react with oxygen and
form polar species such as carboxylic acids which adsorb onto the metal surface and react with the metal
forming metal complexes (Hsu et al., 1988). These metal complexes are soluble in oil, and both homogeneous and heterogeneous catalytic reactions take place. Such reactions generally follow the reaction
mechanism (Hucknell, 1974):
Initiation:
Propagation:
Termination:
© 2001 by CRC Press LLC
Rn M + RH → R• + RH + Rn–1 M
R• + O2 → ROO•
RO• + RH → ROOH + R•
Rn–1MH + ROOH → ROO• + Rn–1MH
Rn–1MH + ROOH → RO• + Rn–1M + H2O
Rn–1MH + ROH → R• + Rn–1M + H2O
Rn–1M + C – O/C = O → RC•O + Rn–1M + H2O
R• + Rn–1MH → Rn–1M + ROH and further oxidation products
R• + RnM → R – R + Rn–2M
ROO• + Rn–1M → Rn–1M – ROO
FIGURE 12.3
Catalytic effect of different metals on oxidation rate.
The role of metal catalysis on lubricant oxidation is a complex one. Since the molecular species in a
lubricant are numerous, the possible reaction pathways and the number of isomers are astronomical.
Because the fundamental reaction mechanisms are functions of the molecular structures, the mechanism
described above can only serve as an illustration of the general directions of the reaction steps. Detailed
understanding of the catalysis mechanism is currently not available.
Different metals exhibit different degrees of catalytic activities on lubricant oxidation. Figure 12.3
illustrates the effects of different metal surfaces on the rate of oxidation (Lahijani et al., 1981). Low carbon
steel has the highest “catalytic” effect in terms of causing the lubricant to oxidize.
12.3.2.4 Polymerization Reactions
Lubricant molecules upon oxidation always tend to go in two different directions: smaller molecules
through beta carbon scission and/or decomposition; high-molecular-weight “polymers” through condensation reactions. Because of the myriad molecular species present in the lubricant molecular mixtures,
there is a statistical averaging effect in terms of product mix and distribution among all lubricants. A
typical aldol condensation reaction is shown below:
O
O
OH
|
CH3–C–CH3 + CH3–C–CH3 → CH3–C–CH3 → CH3–C–CH3 + H2O
|
CH2
CH
|
|
CO
CO
CH3
CH3
The conjugated double bonds of the products are characteristic of the condensation reactions. They have
been experimentally detected by NMR on surface films produced under base oil lubricated contacts
(Naidu et al., 1984). These conjugated double bonds then provide the impetus for further polymerization
to higher-molecular-weight products.
Naidu et al. (1984) demonstrated that the chemistry proceeds in the following sequence: primary
oxidation step; formation of organic acids; aldol condensation reactions to form high-molecular-weight
compounds. When the molecular weight reaches the solubility limit (about 100,000), the reaction products become insoluble and deposit on the surface. Figure 12.4 shows the molecular weight increase as a
function of oxidation time. The lubricant is subjected to thin film oxidation at temperatures of 225 and
© 2001 by CRC Press LLC
FIGURE 12.4
Effect of oxidation on high-molecular-weight reaction product formation.
275°C for different oxidation durations. The surface reaction products are dissolved in a polar solvent,
typically tetrahydrofuran (THF), and analyzed by passing through a gel permeation column for molecular
size separation. The detailed procedure is described elsewhere (Cho and Klaus, 1983).
As shown in Figure 12.4 (Cho and Klaus, 1983), the original lubricant has an average molecular weight
of about 400, which is indicated by the solid zero minute line. As the oxidation continues, the highermolecular-weight fraction increases in magnitude. At a higher temperature of 275°C, the increase in
molecular weight is much faster, but the trend is the same. Similar behavior has been observed under
dynamic wearing conditions, as will be discussed in a later section on organometallic compounds.
As one can see, oxidation reactions are complex, and they are necessary to form the polar species,
which in turn react with the metal surfaces, forming polymers which lubricate the surface. The linkage
between oxidation and wear has long been suspected, but the detailed mechanistic explanation is lacking.
It has long been postulated by the lubricant researchers’ community that more oxidation-resistant oils
are intrinsically more wear resistant. But as we shall demonstrate, the relationship between oxidation
© 2001 by CRC Press LLC
FIGURE 12.5
Effect of oxidation on wear for IIID reference oils.
reactions and film formation tendencies is complex. Figure 12.5 shows the relationship between oxidation
and wear for a set of five reference oils (ASTM sequence III engine dynamometer test for oxidation and
wear; high wear reference oils produce higher wear in the engine dynamometer tests, low wear reference
oils produce lower wear). Four-ball wear test results are measured before and after oil thickening tests
(OTT) on the lubricants (oil thickening tests conditions: 60 mL/min air flow rate, 171°C, 5% drain oil
as catalyst). As can be seen, the high wear reference oils (77B and 77C) have a much higher response to
oxidation than the low wear reference oils.
12.3.3 Organometallic Chemistry and Tribochemistry
In examining the nature of chemical reactions occurring in a contact, the direct reaction between the
surface and lubricant molecules has been alluded to. In this section, we will discuss the role of chemical
reactions induced by mechanical rubbing (tribochemistry).
12.3.3.1 Direct Reactions with Metals
Buckley (1974) has suggested that the nascent surfaces of metals behave differently than oxide-covered
surfaces. Morecroft (1971) performed high-vacuum experiments to show chemical reactions occurred
on freshly exposed metal surfaces. Exoelectrons have been measured and used to explain the enhance
reactivity of the freshly exposed metal surfaces. These early works sparked research into the effects of
nascent surfaces on chemical reactivity. In general, it was found that nascent surfaces under high-vacuum
conditions, readily reacted with any hydrocarbon, decomposing the molecules into fragments. No polymerization reactions have been reported. A conclusion may be that freshly exposed metal surfaces will
readily react, form oxides, and decompose hydrocarbons. In reality, engineering surfaces are almost always
oxide covered. How this observation fits into lubrication theory is not certain.
Hsu and Klaus (1978) used chemical reaction kinetics to back-calculate the reaction temperatures
necessary to generate the observed amount of products from the wear processes and the thermally induced
simulations. The reaction temperatures determined for steel systems are very similar, 379 ± 28°C for the
static case (oxide-covered surfaces) and 349 ± 8°C for the dynamic rubbing case (fresh metal surfaces
exposed). Hsu concluded that, based on these results, the nascent surface effects on reaction rates under
atmospheric conditions are minimal. Thermal effect is the dominant factor for controlling the boundary
chemical reactions for steels.
© 2001 by CRC Press LLC
FIGURE 12.6
Comparison of antiwear effectiveness of different additives with SiC and Si3N4 .
12.3.3.2 Tribochemistry Observed for Ceramics
For ceramics, such as aluminas, silicon nitrides, and silicon carbides, the situation turns out to be different.
Fundamentally, these are polycrystalline solids with very well-defined crystalline structures and bonding
energies. When the surfaces are disrupted, dangling bonds are created. Since these bonds are atom specific,
they are not only more reactive but also tend to be more reaction specific than thermally controlled
reactions. Hydroxide formation was first reported by Gates et al. (1989) for aluminas, and subsequently,
hydration reactions were reported in wearing systems as well as in static systems for other ceramics
(Tomizawa and Fischer, 1986; Mizuhara and Hsu, 1992). Gates and Hsu (1995), in studying alcohols
lubricating silicon nitrides, discovered that the alkoxide reactions dominated the tribochemistry, and
these reactions cannot be simulated by thermal conditions. In studying the surface reactions, Deckman
(1995) compared the reactions of silicon nitride with silicon carbide under identical conditions. Even
though the surface species are predominantly the same, namely silicon oxides and oxynitrides, the
response to different chemistry under rubbing experiments is totally different. Figure 12.6 illustrates the
lack of correlation between the two materials responding to the same chemical compounds under the
same conditions.
While these results are puzzling, they point to the importance of tribochemistry. In metal systems the
thermal reactions dominate and the nascent surface effects are minor. For crystalline materials, not only
is the reactivity different, but the nature of the reaction pathways is different. The cause for this difference
may be suggested by some recent results reported by Nakayama and Hashimoto (1991) and Dickinson
et al. (1993). They independently studied the charged particle emission from crystalline surfaces and
found under deformation and fracture conditions, different materials emit different amounts of particles
(electrons, molecules, charged particles). Figure 12.7 shows that different materials under rubbing conditions emit different amounts of charged particles. The amount of emission for silicon carbide is much
lower than silicon nitride. This observation may not explain directly the difference in reactivity observed
for the two materials, but it does point to a direction for some future research to resolve this issue.
12.4 Boundary Lubricating Films
As early as 1958, Hermance and Egan (1958) noted that the lubrication of metallic surfaces was often
accompanied by opaque organic deposits on and around the contact surfaces. Analysis revealed that these
were high-molecular-weight “polymeric” materials. Over the years, the term “friction polymers” was
coined over the objections of polymer chemists.
The early perception is that this film is responsible for successful lubrication of the surface, and its
formation is controlled by the chemical additives in the lubricants. Therefore, many research efforts have
been focused on the mechanical properties of these films, the nature of them, and in their information
mechanisms.
© 2001 by CRC Press LLC
FIGURE 12.7 Charged particle emission of different surfaces under rubbing conditions. (From Nakayama, K. and
Hashimoto, H. (1991), Triboemission from Various Materials in Atmosphere, Wear, 147 (2):335-343. With permission.)
Recent data suggest that not all films are protective (Hsu, 1991). Depending on the nature of the solid
surfaces, many films can be formed. Some of them are protective, some of them are corrosive, and some
are simply the reaction product residue, which does not affect the friction and wear.
© 2001 by CRC Press LLC
Direct visual observation of formation of ZDP films in a contact during rubbing has provided additional insight into the complex dynamic nature of the process (Sheasby et al., 1991). Similarly, in an
elastohydrodynamic regime, Gunsel and Spikes measured ZDP film thickness as a function of test
duration and temperature (Gunsel et al., 1993; Spikes, 1996).
12.4.1 Physical and Chemical Properties
There have been many studies conducted to analyze the physical and chemical properties of boundary
lubricating films (Belin et al., 1989; Briscoe et al., 1992; Gates et al., 1989; Godfrey, 1962; Hsu and Klaus,
1979; Klaus et al., 1987; Klaus et al., 1985; Lindsay et al., 1993; Martin et al., 1986; Morecroft, 1971; Mori
and Imaizumi, 1988; Tonck et al., 1986).
Results of these studies suggest that the chemical compositions of the films are mainly micron- and
submicron-sized particles of iron and iron oxides intertwined with high-molecular-weight organometallic
compounds of 3000 to 100,000 MW (Gates et al., 1989). If antiwear additives such as zinc dialkyldithiophosphate or tricresylphosphate are present in the lubricants, iron phosphates and phosphate glasses can
be formed and become part of the boundary film (Belin et al., 1989; Martin et al., 1986).
The appearance and morphology of the films can be patchy, continuous, or discrete, and have different
colors, from green to brown to black (Lindsay et al., 1993; Klaus et al., 1987). This is probably due to the
different chemical compositions as well as the thickness of the films. Different thicknesses reflect and refract
part of the light spectrum resulting in different colors being observed (Choa et al., 1994). Overall, there is
no general correlation between the appearance and morphology of the films to effective boundary lubrication.
Several boundary lubricating films are illustrated in Figure 12.8. These are photomicrographs taken
of the wear scars from ball-on-three-flat tests conducted on silicon nitride at 2 GPa mean pressures (Gates
and Hsu, 1995). The films range from fluid-like, observed for 2-ethyl hexyl ZDP (Figure 12.8a) and
calcium phenate (Figure 12.8b), to the more solid-like films of tricresyl phosphate (Figure 12.8c) and
magnesium sulfonate (Figure 12.8d). The film formed from ZDP appears to be flexible, as shown by
Figure 12.9.
There are several mechanisms by which boundary lubricating films function: sacrificial layer, low shear
interlayer, friction modifying layer, shear resistant layer, and load bearing glasses. The sacrificial layer is
based on the fact that the reaction product layer is weakly bound and easily removable; thereby providing
a low shear interfacial layer against the rubbing. So instead of the surfaces being worn by the shear stresses,
this layer is removed instead. For such films to be effective, the rate of film formation has to be higher
than the rate of film removal to protect the surface. An example is the oxide layer in the case of steels
(Quinn et al., 1984). The oxide layer forms rapidly and is easily removed, thus protecting the steel surfaces.
The low shear interlayer mechanism can best be illustrated by the use of solid lubricant molecules. The
solid lubricant molecules have weak interlattice attraction between shear planes; therefore, the lubricating
film can slide easily along the low shear planes within the film, thus accommodating the motion and shear.
Another mechanism is that the molecules or reaction products form an ordered structure at the interface,
and the sliding of the two surfaces is accomplished between the two weakly bonded absorbed layers in the
ordered structure. This is how friction modifiers such as fatty acids function. Of course, the other alternative
mechanism is to have a strongly adhered bonded layer which is shear resistant by itself. Under certain
conditions, the lubricant layer will behave like a solid exhibiting limiting shear and shear band fracture
(Bair et al., 1993). These mechanisms operate in different regimes controlled by the environment and
operating conditions. What works for one system may not work for another. Within the same system,
what works within one set of operating conditions may not work when the operating conditions change
significantly. Figures 12.10 and 12.11 illustrate different film morphologies and appearances for two
different materials. Figure 12.10 shows very thin dense films that work for silicon nitride. These films are
similar in appearances and textures to the steel-on-steel systems. Figure 12.11 shows effective films for the
silicon carbide system. The same chemicals in Figure 12.10 do not work for silicon carbide. Because of the
brittleness of silicon carbide, thicker films are needed to protect the system.
© 2001 by CRC Press LLC
FIGURE 12.8 Optical micrographs of wear scars from wear tests on silicon nitride using paraffin oil containing
(A) 1% 2-ethylhexyl ZDP; (B) 1% Ca phenate. Optical micrographs of wear scars from wear tests on silicon nitride
using paraffin oil containing (C) 1% tricresyl phosphate; (D) 1% Mg sulfonate.
FIGURE 12.9 SEM micrographs of film from wear tests on silicon nitride using paraffin oil containing 1%
2-ethylhexyl ZDP.
© 2001 by CRC Press LLC
FIGURE 12.10 SEM photomicrographs of wear scars from wear tests on silicon nitride using paraffin oil containing
1% 2-ethylhexyl ZDP.
12.4.2
The Detection of Organometallic Compounds in Films
Oil-soluble metal-containing compounds were first identified as being generated in lubricants under
oxidizing conditions (Klaus and Tewksbury, 1973). These compounds were later identified to be highmolecular-weight organometallic compounds using gel permeation chromatography coupled with atomic
absorption spectroscopy (Gates et al., 1989). Figures 12.12 through 12.15 show the results for two cases:
static oxidation test conditions and dynamic wear test conditions. Under static oxidation test conditions,
a thin oil film about 40 µm thick was deposited on a steel disk and oxidized at a temperature of 225°C
for 20 to 30 minutes. Afterwards, the surface reaction products were dissolved by a solvent (tetrahydrofuran). The solution was then injected into the GPC columns for molecular size separation. The effluent
flowed through two detectors, refractive index and ultraviolet detector. After the detectors, the effluent
stream was collected in a autosampler vial for the determination of metal content in the effluent stream
by atomic absorption spectroscopy analysis. The same procedures were followed to examine the films
formed on worn surfaces after wear experiments.
Figure 12.12 shows that organometallic compounds are formed when lubricants are reacted with the
metal surfaces under static oxidation conditions. The top curves denote the molecular weight distributions of the lubricant (a 150 N paraffinic base oil) after 30 minutes of reaction time at 225°C on steel
and copper surfaces. The original molecular weight distribution centers around 300 (similar to the
molecular weight distribution curve for the copper surface). The lower two curves are the results of the
atomic absorption signals as a function of the discrete volumetric increments of the effluent from the
© 2001 by CRC Press LLC
FIGURE 12.11 SEM photomicrographs of wear scars from wear tests on silicon carbide using paraffin oil containing
1% benzyl phenyl sulfide.
gel permeation column. So these signals indicate the amount of organometallic compounds as a function
of molecular weight. In these two cases, one can see that the steel surface, under identical reaction
conditions, forms a larger quantity of high-molecular-weight products and produces a large quantity of
organo-iron compounds. The copper surface, on the other hand, produces a much smaller amount of
organo-copper compounds (note the scale difference between steel and copper) even though the molecular weights of the organometallic compounds are about the same. The molecular weight distribution
curves also suggest that copper did not cause the original molecular weight distribution to change
substantially, i.e., the lubricant is not significantly oxidized.
Figure 12.13 shows the formation of organo-iron compounds as a function of time under static
oxidizing conditions. As one can see, the iron peak intensity increases with time. This indicates that the
amount of organo-iron compounds are increasing as a function of time. The data also indicate that there
is an induction time under static oxidation conditions for the organo-iron compounds to form. The
analytical procedures used here are sensitive to ppm level of iron, and the amount of organometallic
compounds detected depends on the solvent and the extraction procedure.
Figure 12.14 shows the result when the same procedure is applied to the dynamic wear case for a superrefined mineral oil base stock in a four-ball wear tester. The wear procedure used is a modified procedure
in that only 6 µL of lubricant is available to the contacts. This way, the reaction products and the reaction
sequence are concentrated for ease of analysis. A broad spectrum of organo-iron compounds of various
molecular weights is found, with molecular weights ranging up to about 100,000. Higher-molecularweight compounds are not detected, suggesting that the solubility limit has been reached Figure 12.15
shows the results for a fully formulated lubricant (containing antiwear additives, detergents, dispersants,
etc.) as a function of time. In this case, the amount of organo-iron compounds is much less, but the
presence of organo-iron compounds can be detected very rapidly. Optical pictures reveal that the boundary
© 2001 by CRC Press LLC
FIGURE 12.12 Molecular weight distribution of organometallic compounds from oxidation tests conducted on
iron and copper surfaces.
lubricating film is fully formed after only 1 minute of wearing contact. There seems to be a shearing
action reducing the molecular weight of the products and shifting the maximum amount of organo-iron
to a molecular weight of about 3000.
These organometallic compounds are also found on cam and tappet parts in an ASTM engine dynamometer test, the sequence III oxidation wear test. Cam and tappet parts were taken from ASTM test
stand calibration runs and analyzed for surface reaction products. Similar patterns were observed. This
suggests that organometallic compounds play an important role in the formation of the boundary
lubricating films.
12.4.3 Mechanical Properties of Boundary Lubricating Films
Let us hypothesize that these high-molecular-weight organometallic compounds form the glue to provide
cohesive strength with dispersed fine particles of iron and iron oxides. The bonding via the iron organometallic bonds provides the adhesive strength. Then once the film is formed, one should be able to
measure the shear strength of these films.
In the last few decades, many researchers have attempted to measure the film strength. There have
been some limited successes. Briscoe et al. (1973) studied the shear strength of thin films under a range
of high pressure (MPa to GPa). They found that at these pressures shear strengths increased with
increasing pressure for calcium and copper stearate and polyethylene films. Increasing temperature
reduced shear strength. Researchers at École Centrale de Lyon have devised clever ways to measure the
shear strength of thin films by a microslip method (Tonck et al., 1986). These strengths are tabulated in
Table 12.1. Generally, the film strength increases with load for base fluids, zinc dithiophosphates, and
© 2001 by CRC Press LLC
FIGURE 12.13
at 250°C.
Molecular weight distribution of organometallic compounds from static oxidation tests conducted
calcium sulfonates but decreases with load in the case of a friction modifier. This is reasonable since
friction modifiers depend on multilayer adsorption and easily sheared planes between the molecular
layers. As load increases, the film thickness will decrease. Monomolecular film studies have also been
done (Briscoe and Evans, 1982). The measured shear strength for stearic acid was found to be on the
order of about 10 MPa on a steel surface.
Comparison of these two results is difficult because of the differences in measurement techniques, film
chemistry, and film thickness. Yet in terms of an order of magnitude comparison, it is instructive to note
that a monolayer is relatively weak and the complex films generated by lubricants are relatively strong.
Another observation is that not all films lubricate (Deckman, 1995). Hsu (1991) suggested that there is
an optimum reactivity for a film to lubricate the surfaces based on the commonly observed constant
renewal, sacrificial lubrication mechanisms. The ability of the film to lubricate the surfaces depends on
the adhesive and cohesive strengths of the film with respect to the surface.
Warren et al. (1998) used scanning force microscopy to determine nanomechanical properties of ZDP
films from both static (heating) and wear tests. They found that films generated in the highly loaded
regions of the wear test were markedly different in nanomechanical response than films from a lightly
loaded region and films from a static test.
We have attempted to show the linkage among the oxidation reactions, polymerization reactions, and
the effects of different materials in effecting lubrication. The picture that is emerging is complex but
traceable to molecular phenomena on the surface. The surface is an integral part of the reacting system.
Since surfaces depend on many variables including machining, defects, crystalline configurations, oxide
layers, hydride layers, etc., the resulting reaction pathways sometimes are difficult to predict. But knowing
© 2001 by CRC Press LLC
Intensity, arb. units
Apparatus: 4-ball wear tester
Speed:
600 rpm
Load:
40 kg
Atmosphere: 0.25 l/min dry air
Lubricant: 6ml superrefined mineral oil
Duration: 30 minutes
Material: 52100 steel
Molecular weight distribution
(refractive index detector)
Fe peak intensities
106
105
104
103
10
2
101
10
0
Molecular Weight
FIGURE 12.14
Molecular weight distribution of organometallic compounds from dynamic wear tests.
the surface layer composition precisely, and the operating conditions, the future of predicting boundary
lubrication is foreseeable, albeit with the existence of many formidable obstacles.
With the many studies on boundary film formation, some generalizations can also be made.
Figure 12.16 shows the general diagram for the formation of boundary films.
The area of tribochemistry needs many more studies to understand the intricate interactions between
surfaces and molecules under rubbing conditions. The role of surface defects as well as dangling bonds
(Lenahan and Curry, 1990) needs to be identified and characterized in order to gain important insights
into surface reactions which have wide implications for other areas of research as well.
12.4.4 Advances in Measurement Techniques
There are several new techniques that have been applied to boundary lubrication issues, ranging from
new wear and property measurement to specialized surface and film analysis. In general, these new
techniques have followed a trend toward smaller-scale as well as more specific property measurement.
At the small end of the scale — atomic-level property measurement — the surface forces apparatus
(SFA), pioneered by Israelachvili (Israelachvili, 1986), is being used to probe the properties of the first
monolayer of molecules on surfaces. Significant progress has been made on the properties of confined
films, even though the technique is limited to transparent, atomically smooth, materials like mica.
Scanning probe microscopy (SPM), first introduced as a technique for looking at the topography of
surfaces at the nanometer level, is also now being used essentially as a “nanotribometer.” Many different
forms of the nanotribometer SPM have been used to measure friction, adhesion, and wear at this ultrasmall scale (Mate, 1995; Feldman et al., 1998; Bhushan et al., 1995). At this scale however, issues of
calibration and control are critical and continuously evolving, with recent designs for instruments such
as the NIST “calibrated” atomic force microscope (Schneir et al., 1994) as well as commercial “metrology”
instruments making their way into use. The issue of cantilever calibration, so critical to obtaining accurate
absolute values for forces is being approached from several angles such as finite element analysis (Hazel
© 2001 by CRC Press LLC
FIGURE 12.15
Molecular weight distribution of organometallic compounds from wear tests of different duration.
TABLE 12.1 Mechanical Properties of Surface Films Measured by Microslip
Technique: Cast Iron on 52100 Steel
Measure Shear Moduli at Selected Normal Load
(GPa)
Film
1N
2N
5N
10N
Film A, 120 nm dodecane
Film b, 60 nm ZDP
Film C, 60 nm Ca sulfonate 5%
Film D, FM complex ester 1%
3
3.5
1.9
3.4
3.3
3.8
2.1
2.9
3.6
4.7
3.1
2.5
4.3
4.2
3.2
1.8
Data from Tonck, A., Kapsa, Ph., and Sabot, J. (1986), Mechanical behavior of
tribochemical films under a cyclic tangential load in a ball flat contact, J. Tribology,
Vol. 108, 117.
and Tsukruk, 1998), resonance methods (Cleveland et al., 1993), and reference cantilevers (Gibson et al.,
1996). A simple, accurate, calibration for both normal and tangential forces is still elusive.
The issue of relevance of these nanoscale measurements to larger “real world” devices is still being
debated. Interestingly enough, the relevance issue is being bridged by the fact that actual devices are now
shrinking down to the scale of measurement — as in the rapidly developing fields of micromachines and
microelectromechanical systems (MEMS).
Some very interesting contributions are being made on slightly larger scales with modified SFA and
SPM apparatus. These tribometers and indenters make very small-scale measurements of surfaces and
films in the nanometer and micrometer scale contact regime. One of the simplest SPM modifications,
© 2001 by CRC Press LLC
FIGURE 12.16
Overall framework for boundary lubricating film formation.
first made by Ducker (Ducker et al., 1991), was the attachment of very small spheres to the end of an
SPM cantilever. Instead of a contact with 10 nm radius of contact that might just push a film aside, the
contact could be enlarged to micrometer dimensions. In another approach, the research group at L’École
Central (Bec et al., 1996; Tonck et al., 1986; Georges et al., 1998) have developed an apparatus that utilizes
a variety of diamond and sphere tips to apply controlled forces to surfaces and measure their effects.
They have made considerable progress in probing the properties of important boundary lubricating thin
films such as ZDP on surfaces.
In boundary lubrication research, the key has always been unlocking the understanding of the underlying chemistry taking place during lubrication. In many cases this means using the combination of wear
testing and key analytical instrumentation designed to elucidate the chemical mechanism responsible for
lubrication. In the area of analytical measurements of surface films, the workhorse techniques of FTIR,
Raman, XPS, and Auger are being augmented with highly sensitive and specific techniques such as near
edge X-ray absorption fine structure (NEXAFS) spectroscopy, which utilizes a synchrotron radiation
source to generate soft X-ray beams to measure the presence and orientation of molecularly thin films
on surfaces. This technique was first applied to stearic acid on copper and was able to provide a detailed
understanding of the surface bonding under both static and dynamic rubbing conditions (Fischer et al.,
1997a,b). The technique was also recently applied to nanometer-thin layers of perfluoropolyether
© 2001 by CRC Press LLC
containing phosphazene on magnetic hard disks to show the presence of unsaturated carbon bonds at
the interface (Kang et al., 1999). Time of flight secondary ion mass spectrometry (TOF SIMS) has been
used to probe the chemical composition of relatively high-molecular-weight films formed from silicon
nitride lubricated by a monoalcohol (Gates and Hsu, 1995). The technique showed unambiguously that
silicon alkoxides were being produced by tribochemical reactions during rubbing and helped to define
the tribochemical mechanism of boundary lubrication of silicon nitride by alcohols.
Other techniques such as microellipsometry and surface reflectivity are being utilized in direct response
to specific industrial needs. In the case of the hard disk industry, surface reflectivity techniques are being
refined in such a way that depletion of the nanometer-thin lubricant film and wear of the protective
carbon overcoat layer can be measured in situ during component testing (Meeks et al., 1995).
12.5 Boundary Lubrication Modeling
Boundary lubrication effectiveness has long been considered to be essential in modern machine designs
for proper operation. As the demands for better energy efficiency, tighter tolerances, and the availability
of new materials, the need to predict lubrication effectiveness in this regime increases.
What do we mean by effective predictive models? Given the materials pair, speed, load, surface
roughness, lubricant type (viscosity, additive chemistry), and duty cycles, can we predict length of service,
amount of wear, time to scuffing, seizure? According to this definition, we currently do not have any
such model. At the same time, we can describe fairly well average film thickness, elastohydrodynamic
support, and even some wear.
It is instructive, therefore, to examine the current predictive ability on some of these parameters: wear,
flash temperatures at the tips of the asperities, lubrication transition temperatures, and some discussion
on the subject of molecular dynamic simulation.
12.5.1 Wear
Wear is a system function. It depends on the system, which includes the materials, surface roughness,
lubricants, environment, operating conditions, temperatures, etc. To predict wear a priori, i.e., without
experimental fitting constants, is very difficult. Yet wear is such an important parameter, so studies abound
to describe the wear processes (Choa et al., 1994).
Under boundary lubrication regime, one of the most critical initial system parameters is the real area
of contact. This parameter controls the real load the asperities are under and the subsequent stress/strain
relationship beneath the contact. As discussed in the chapter titled, “Wear Maps,” the fundamental wear
process for most metals is controlled by the accumulation of strain. Greenwood and Williamson (1966)
proposed several models to describe the initial real area of contact, which is usually a small fraction of
the apparent area of contact between two engineering surfaces. One version, which assumes a distribution
of peak heights can be written as:
()
A = π ηβσ F1 h
where η = number of asperities
β = asperity radius
σ = standard deviation of the peak height distributions
and
∞
( ) ∫ (s − h) φ * (s)ds
Fm h =
h
where φ* (s) is a probability density.
© 2001 by CRC Press LLC
m
However, once the sliding takes place, the real area of contact changes with wear. Under steady-state
mild wear condition the real contact area could be quite high when the two surfaces conform to one
another.
The next step is to model the lubricant rheology in the contact. Under boundary lubrication conditions,
traditionally, bulk fluid viscosity does not play a role since the asperities are bearing the load. However,
recent studies of fluid molecules under confined space exhibit radically different mechanical and flow
properties as compared to those in the bulk (Coy, 1998; Granick, 1991a,b; Granick, 1999; Israelachvili,
1993, 1986; Israelachvili et al., 1988). The viscosity near the wall (nanometers) is much higher than those
of the bulk. This also implies that viscosity of the fluid trapped inside the contact among the asperities
under sliding conditions may be higher, hence giving rise to the elastohydrodynamic life at the local level.
Some of this viscosity and polymer solution behavior near the wall can be simulated by dissipative particle
dynamics (DPD) (Hoogerbrugge and Koelman Jmva, 1992; Kong et al., 1997), which is an accelerated
molecules dynamics method.
Once the rheology is defined, fluid film thickness can be estimated by Dowson and Higginson’s
(Dowson and Higginson, 1959) equation:
 η0.7 α 0.54 V 0.7 R0.43 
h = 1.63 0 00.13

L
E ′ 0.03


There are several computer programs available to calculate the contact stresses, fluid flow, temperatures,
and elastohydrodynamic lifts (Lee and Cheng, 1992), but to link these calculations to wear is a big step.
Bell (Bell and Colgan, 1991; Bell and Willemse, 1998) used component cam-follower wear data to correlate
with oil film thickness and calculate the oil film thickness using a finite element program. The wear
contours and approximate amount of wear of the cam-follower contact in an engine was successfully
simulated. Chemical film formation as a function of additive concentrations and temperatures was not
taken into account but simulated by the bench tests.
12.5.2 Flash Temperatures
Engineering surfaces are not atomically smooth. Under contact conditions, the hills (asperities) of each
surface bump into the hills of the other surface. When they move relative to one another, frictional energy
producers heat, and the transient temperatures at the tips of asperities are called flash temperatures. They
are a microsecond in duration and highly localized (Furey, 1964).
When one surface begins to move over another surface, the contact phenomena are controlled by
asperity interactions and the interfacial layer (liquid lubricant or oxide). At very light loads, the friction
is controlled by the surface forces which include van der Waals forces, hydration force, electrostatic or
double layer forces (depending on the materials), and elastic contacts of the asperities and the viscous
drag of the oil film. The elastic contacts will sometimes produce chatter, or stick slip phenomena as
observed by frictional force traces.
Under high load, the asperity contacts result in plastic deformation of the asperities. This changes the
surface features of the interface. The asperity–asperity contacts lead to deformation of the asperities, and
this process provides the majority of the frictional resistance. Other processes contributing to friction
and heat release are adhesion, plowing, and abrasion.
In boundary lubrication, chemical films are produced to provide surface protection. The films for steel
systems are produced mostly by frictional heating. Therefore, if one is able to predict the asperity
temperatures in the contact one can calculate the chemical reaction rate to generate the films.
Historically, two approaches have been used to estimate the flash temperature. One is chemical; the
total amount of reaction products is measured and compared to constant temperature reaction rates at
different temperatures. Given the time and temperatures, from the total reaction products, the average
surface reaction temperatures can be estimated. The other approach is mechanical; asperity contacts are
© 2001 by CRC Press LLC
modeled to take into account the elasticity, frictional heating, fluid flow, heat transfer characteristics as
a function of load, speed, surface roughness, and materials properties.
To compare these two approaches, Hsu, Klaus, and Cheng (1988) used a simple experiment on a fourball wear tester. The chemistry of the lubricant was kept simple by using a purified paraffinic base oil.
A microsample test procedure was used to allow maximum lubricant degradation. The composition of
the lubricant during and after the test was analyzed to show molecular weight changes. The total amount
of reaction products generated during the wear test was measured. Knowing the kinetics of the reaction,
the temperatures in the contact were calculated. The temperatures were also determined by using a
mechanical contact model given the surface roughness, material properties, and physical properties of
the lubricant. These two temperatures were then compared.
Klaus et al. (1980) demonstrated that conditions within a thin-film micro-oxidation test closely resemble those within a sliding contact. The test used a thin lubricant film on a metal catalyst cup in a
nitrogen–oxygen mixture, thus both temperature and oxygen availability may be controlled. Subsequent
work (Clark et al., 1985; Klaus et al., 1982) demonstrated that the oxidation reactions followed the
Arrhenius relationship. The equation has the form:
k = m e − E RT
where T = temperature
m = a fitting coefficient
E = the activation energy (cal/mode)
R = gas constant
Naidu et al. (1986) described a global rate model based on the consumption of lubricant within a microoxidation test, subjected to conditions similar to those encountered in a sliding contact. The model
describes the primary oxidation step as well as the subsequent condensation polymerization step, which
results in lubricant viscosity increase and insoluble sludge formation. These reactions can be described
as follows:
where A
E
B
F
P
D
k
= original oil
= evaporated original oil
= low-molecular-weight oxidation products
= evaporated low-molecular-weight products
= high-molecular-weight liquid polymerization products
= insoluble deposits
= reaction rate constants
The estimated contact temperature required to produce polymerization in 10% of the total lubricant
volume is shown in Figure 12.17, as a function of film thickness. As the mean separation between the
opposing surfaces decreases, the theoretical oil temperature increases exponentially, as less oil is able to
enter the contact junction.
The mean boundary film thickness between the opposing surfaces is approximately 0.06 µm (600 Å),
as calculated from the asperity contact model detailed in the following section. The required film
temperature at this separation is 375°C. Figure 12.18 shows the calculated lubricant life at this temperature
© 2001 by CRC Press LLC
Estimated Temperature, °C
Film Thickness, µm
FIGURE 12.17
Estimated lubricant temperature as a function of film thickness.
400
Time to Failure, minutes
600 rpm
40 kg
300
3 75
200
°C
100
0
0
1
2
3
4
5
Lubricant Volume, µl
FIGURE 12.18 Calculated oil film temperature required to provide the lubricant oxidation measured during
microsample wear tests with different volumes of lubricant.
as a function of lubricant volume, along with experimentally determined values for comparison. As may
be seen, the predicted values are in general agreement with the experimental data.
The asperity temperatures within the contact zone of two rough surfaces were calculated using the
model developed by Lee and Cheng (1992). The input data include two groups: (1) experimental
conditions such as sample geometries, load, speed, the material’s elastic constants, and lubricant properties, and (2) measured data such as friction and surface roughness profiles of the worn samples. Given
roughness profiles of the opposing surfaces, the calculations include three sequential steps: (1) contact
modeling to determine the relationship between average contact pressure and the average distance (gap)
between the two surfaces, (2) rough-EHD analysis to calculate the average film thickness and the load
distribution between hydrodynamic film support and asperity contacts, and (3) temperature calculation
to determine the asperity temperature distribution within the contact zone.
© 2001 by CRC Press LLC
Surface Roughness
Top Ball
(inverted)
RMS: 0.19
Lower Ball
RMS: 0.31
Calculation Region
Relative
RMS: 0.23
Test: 2160R
Load: 40kg
Scale, µm
Speed: 600rpm
Time: 60 minutes
1
Lubricant: Paraffin Oil
Material: 52100 steel
100
Measurements taken perpendicular to direction of sliding
FIGURE 12.19
Profiles taken from opposing worn surfaces.
To model two surfaces in contact, one has to describe the surface roughness and simulate the contact
geometry under static load mathematically. The contact model used assumes that both surfaces have
longitudinal roughness. The validity of this assumption has been demonstrated (Sugimura and Kimura,
1984; Wang et al., 1991) for wear scars formed after a four-ball wear test. Figure 12.19 shows the surface
roughness profiles taken from the top ball (inverted) and the lower balls after the prewear test. These
profiles were measured perpendicular to the sliding direction. To ensure accurate representation of the
wear scar on the top ball, its roughness trace was determined by averaging digitized profiles in three
angular directions (θ = 0°, θ = 120°, θ = 240°). An average trace for the lower balls was similarly
determined, one profile being taken from each of the lower balls. Subsequently, the profiles from the top
ball and the lower balls were derived, by subtraction, to form a composite profile, denoted as “relative”
trace in Figure 12.19. The contact between the top ball and the lower balls was then simulated by analyzing
the contact between the composite profile and a rigid plane. Taking into account the elastic deformation,
the local high spots (asperities) of the composite profile would deform and become flattened when the
profile is pressed against the rigid plane. The distance measured from the rigid plane to the mean line
(plane) of the deformed composite profile defines the average gap between the mean lines (planes) of the
two rough surfaces under the same load. Meanwhile, the average pressure of those deformed asperities
defines the average contact pressure. By increasing the severity of loading, the average gap becomes smaller,
while the contact pressures of the individual asperities increase. Depending on the local geometries, some
of the asperities may yield as the contact severity increases. When this happens, the model assumes that
the contact pressure stays at the yield pressure and does not increase further. If one varies the loading
mathematically, one could determine the relationship between the average contact pressure and the average
gap, as illustrated in Figure 12.20. For this particular simulation, since the balls are preworn, the maximum
Hertzian pressure under an applied load of 40 kg is estimated to be 0.847 GPa. The average gap in the
central portion of the contact is about 0.06 µm, according to Figure 12.20. The r.m.s. value of the “relative”
roughness shown in Figure 12.19 is 0.23 µm. After the relationship between average contact pressure and
average gap has been determined, the rough-EHD analysis was then carried out.
In the rough-EHD analysis, a line-contact between two rough cylinders is assumed, and only the inlet
half of the contact is analyzed (Lee and Cheng, 1992). The pressure along the centerline of the line contact
is assumed to be the same as the maximum Hertzian pressure determined by considering the preworn
geometry. The film thickness solution of the rough-EHD analysis is based on the average gap. Since the
analysis is based on two rough surfaces in contact, the two cylinders are subjected to two types of pressures;
one from hydrodynamic lift, the other from asperity contact pressures. The relationship between average
contact pressure and average gap determined in the contact simulation previously is therefore utilized
© 2001 by CRC Press LLC
6
Average Pressure, GPa
5
4
3
2
1
0
0.0
0.1
0.3
0.2
Average Gap, µm
FIGURE 12.20
Average contact pressure as a function of gap between mean lines of opposing surfaces.
1.0
Fraction of Centerline Pressure
0.8
Asperity Pressure
0.6
0.4
0.2
Hydrodynamic Pressure
0.0
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
X/b
FIGURE 12.21
Normal pressure as a function of normalized distance from contact center.
in the calculation of the overall pressure distribution of the contact. Figure 12.21 illustrates the loadsharing characteristics between asperity contact and hydrodynamic lift, as a function of the normalized
distance from the center of the contact. The parameter b is the Hertzian contact radius. Clearly, asperity
contacts carry most of the normal load, thus, sliding friction between the opposing asperity contacts will
be responsible for most of the heat generated. Besides the pressure estimations, the rough-EHD analysis
also calculates the average friction coefficient from the asperity contacts based on the measured friction
value (Lee and Cheng, 1992). In this case, the measured friction coefficient was 0.11, and the average
asperity contact friction was determined to be 0.112.
From the rough-EHD calculations, the size and shape of the asperities within the contact zone can be
determined. Each asperity contact forms roughly an elongated ellipse, with its major axis orientated in
the sliding direction. Also, the pressure distribution on each asperity as a function of distance from the
contact centerline can be determined. Before the temperature calculations are performed based on Jaeger’s
work (Jaeger, 1942), some further simplifications are taken. Each asperity contact is approximated to be
rectangular, and its average contact pressure is used for temperature calculations. Frictional heating is
© 2001 by CRC Press LLC
200
Temperature, °C
180
160
140
120
100
0.0
0.1
0.2
0.3
0.4
0.5
Contact Area Ratio
FIGURE 12.22
Cumulative temperature distribution within the Hertzian contact.
considered as the heat source, which includes both the average asperity contact friction and the hydrodynamic film friction. The resulting temperature within each asperity contact is a function of the location
measured from the centerline of the nominal contact zone. The maximum temperature of each asperity
contact is sought and will be considered as the representative temperature for that asperity. Figure 12.22
shows the cumulative temperature distribution under the conditions included in Figure 12.19. The
highest temperature attainable is approximately 220°C covering about 3% of the nominal area. These
results also reveal that the asperity contacts cover approximately 45% of the nominal area, and the bulk
temperature is ~150°C.
The results of the temperature analysis show that the temperatures within the contact zone are in the
range of 150 to 220°C. This is in general agreement with other continuum models (Francis, 1970; Fein,
1960) which, for the present test conditions, predict the maximum contact temperature to be 157 and
160°C, respectively. By using the current model, the considerations of asperity contacts, thus local high
contact pressures, and a slightly higher average friction than the measured overall friction appear to have
already increased the temperature estimates. However, the reaction kinetics model estimates an average
temperature of 375°C in the lubricant film is necessary to degrade 10% of the total lubricant. Also,
measurements carried out by Bos (Bos, 1975) under slightly more severe conditions using a subsurface
thermocouple, demonstrated that the true surface temperatures during a four-ball wear test are in excess
of 300°C. Clearly, the predicted contact temperatures based on the current mechanical model are considerably lower than that required by the chemical reaction model. The discrepancy is large enough to
be significant. For a more detailed discussion on the subject, see reference Hsu et al., 1994.
12.5.3 Asperity-Asperity Understanding
One can effectively argue that most wear events under effective lubrication conditions are occurring at
the asperity level. Conversely, lubrication means protecting the asperities. When asperities are touching
and sliding over each other, both mechanical events and chemical events occur, but the wear is dominated
by the mechanical events in the form of contact pressures, stresses, and strain accumulation rate.
If the load is so high that fracture of the asperity is inevitable, lubrication has little impact. For lower
loads, the protective mechanisms are: load-bearing interface; easily shearable layer; interfilm shear; and
adhesion barrier. Under high pressure, the lubricant or the chemical film behaves like a solid between
the contact, enlarging the real area of contact, hence reducing the contact pressure, redistributing the
stresses, and therefore reducing the strain buildup rate for metals or the stress intensity for brittle solids.
Since wear of brittle solids depends primarily on stress intensity, lowering it minimizes wear. If the film
© 2001 by CRC Press LLC
is easily shearable and has sufficient thickness, then the magnitude of the tensile stresses imposed by the
asperity will be significantly reduced. If the film is thick enough and can allow interfilm shear to take
place easily, such as solid lubricating films with weak planar attractive forces, the relative motion can be
accommodated by interfilm movement. If the molecules are large, have strong bonding to the surface to
resist shear, and cover the asperity surface reasonably well, the presence of such film will prevent nascent
surface contact forming cold welding or adhesion. Under current boundary lubrication technology, many
of these protective mechanisms are being used to protect the surfaces.
12.5.4 Molecular Dynamics Modeling
In the last decade, molecular dynamics modeling of contacts and friction between materials has been
used to understand the atomic and molecular origin of friction and wear (Harrison et al., 1995; Perry
and Harrison, 1996; Stuart and Harrison, 1999). Concurrently, experimental measurements at the atomic
and molecular level have been conducted (Granick, 1991a,b; Granick, 1999; Krim, 1996; Mate, 1995).
These results are providing insights into the fundamental processes of friction and wear.
12.5.4.1 The Issue of Scale
While the molecular dynamic calculations and the atomic scale measurements are providing valuable
insights into the basic processes, the issue of scale emerges. Proponents of atomic scale investigations
point out that if we understand the friction at the molecular level, then we can apply this knowledge to
the micro- and macroscale events and significantly improve the technology. At this time, this has not
taken place.
Let us examine the issue from an engineering point of view. On each asperity, there are subasperities,
which are smaller in scale. On each subasperity, there are sub-subasperities, and so on. At what scale
should the contact of two surfaces be considered? For most engineering applications, the critical scale
for friction is at the micron scale. The sub- subasperities need not be considered. Can one describe fully
the asperity deformation process from the deformation of subasperities, sub-subasperities … down to
the molecular or atomic level? The asperity deformation process is a complex system containing many
stress domains, defects, and various dislocation zones. Within each zone, there are many energy levels
existing at the molecular level. It is not clear how atomic simulations can reach the proper scale factor
to simulate what is happening at the interface. But if a constitutive relationship can be developed from
the molecular dynamic models, it may help to explain phenomena difficult to observe experimentally.
12.5.4.2 Molecular Dynamics Models
The basic calculation procedure has been laid out (Harrison et al., 1993). Take a collection of atoms or
molecules, calculate the interatomic forces, integrate equations of motion within a boundary from an
initial position, according to the temperature, pressure, stress, and energy transfer mechanisms specified
by the model. Periodic boundaries are used to simulate an extended array of atoms. Reactions and
molecular attachment are dictated by energy levels and pairwise potentials (Brenner, 1990).
An example of this technique is shown in Figure 12.23 for the simulated tribochemical wear of diamond
(Harrison and Brenner, 1994). As the model diamond surfaces pass one another under contact (from
a → d), we observe C–C bond formation (adhesion), and subsequent bond breakage resulting in the
transfer of atoms from one surface to the other.
This area of study is rapidly evolving to handle complex situations more relevant to boundary lubrication; however, the current significance to practical lubrication is not clear at this time.
12.6
Concluding Remarks
Significant progress has been made in the field of boundary lubrication research in the past 20 years.
Much of this is focused on understanding the factors that influence lubrication effectiveness. We can
identify the effects of tribochemistry, lubrication film formation, and the failure mechanisms of boundary
© 2001 by CRC Press LLC
FIGURE 12.23 Molecular dynamics model simulation of tribochemical interaction between diamond surfaces.
(From Harrison, J.A. and Brenner, D.W. (1994), Simulated tribochemistry — an atomic-scale view of the wear of
diamond, Journal of the American Chemical Society, 116 (23):10399-10402. With permission.)
lubrication much better; however, our ability to predict lubrication effectiveness in any given instance
for a particular system still needs development.
References
Bair, S., Qureshi, F., and Winer, W.O. (1993), Observations of shear localization in liquid lubricants under
pressure, Journal of Tribology — Transactions of the ASME, 115 (3):507-514.
Bec, S., Tonck, A., Georges, J.M., Georges, E., and Loubet, J.L. (1996), Improvements in the indentation
method with a surface force apparatus, Philosophical Magazine a — Physics of Condensed Matter
Structure Defects and Mechanical Properties, 74 (5):1061-1072.
Belin, M., Martin, J.M., and Mansot, J.L. (1989), Role of iron in the amorphization process in frictioninduced phosphate glasses, Tribology Trans., 32:410.
Bell, J.C. and Colgan, T. (1991), Pivoted-follower valve train wear — criteria and modeling, Lubrication
Engineering, 47 (2):114-120.
Bell, J.C. and Willemse, P.J. (1998), Mid-life scuffing failure in automotive cam-follower contacts, Proceedings of the Institution of Mechanical Engineers Part J — Journal of Engineering Tribology, 212
(J4):259-269.
Bhushan, B., Israelachvili, J.N., and Landman, U. (1995), Nanotribology — friction, wear and lubrication
at the atomic scale, Nature, 374 (6523):607-616.
Blencoe, K.A., Roper, G.W., and Williams, J. (1998), The influence of lubricant rheology and surface
topography in modelling friction at concentrated contacts, Proceedings of the Institution of Mechanical Engineers Part J — Journal of Engineering Tribology, 212 (J6):391- 400.
Booser, E.R. (Ed.) (1984), Handbook of Lubrication, CRC Press: Boca Raton, FL.
Bos, A. (1975), The temperature at the wear scars of the four-ball apparatus, Wear, Vol. 31, 17.
Brenner, D.W. (1990), Empirical potential for hydrocarbons for use in simulating the chemical vapordeposition of diamond films, Physical Review B — Condensed Matter, 42 (15):9458-9471.
Briscoe, B.J. and Evans, D.C.B. (1982), The shear properties of Langmuir–Blodgett layers, Proc. R. Soc.
Lond. A Vol. 380, 389.
Briscoe, B.J., Scrunton, B., and Willis, F.R. (1973), The shear strength of thin lubricant films, Proceedings
of the Royal Society of London Series a — Mathematical Physical and Engineering Sciences, 333:99-114.
Briscoe, B.J., Thomas, P.S., and Williams, D.R. (1992), Microscopic origins of the interface friction of
organic films: the potential of vibrational spectroscopy, Wear, Vol. 153, 263.
© 2001 by CRC Press LLC
Buckley, D. (1974), Friction induced surface activity of some simple organic chlorides and hydrocarbons
with iron, ASLE Trans., Vol. 17, 36.
Buckley, D.H. (1981), Surface Effects in Adhesion, Friction, Wear, and Lubrication, Elsevier: New York.
Cheng, H.S. and Lee, S.I. (1989), Modeling of thin-film lubrication breakdown in sliding contacts,
Abstracts of Papers of the American Chemical Society, 197:108-COLL.
Cho, L.F. and Klaus, E.E. (1983), Microoxidation evaluation of automotive crankcase oils, SAE paper
831679, Society of Automotive Engineers, Warrendale, PA.
Choa, S.H., Ludema, K.C., Potter, G.E., DeKoven, B.M., Morgan, T.A., and Kar, K.K. (1994), A model of
the dynamics of boundary film formation, Wear, Vol. 177, 33.
Clark, D.B., Klaus, E.E., and Hsu, S.M. (1985), The role of iron and copper in the oxidation degradation
of lubricating oils, Lubrication Engineering, 41 (5):280-287.
Cleveland, J.P., Manne, S., Bocek, D., and Hansma, P.K. (1993), A nondestructive method for determining
the spring constant of cantilevers for scanning force microscopy, Review of Scientific Instruments,
64 (2):403-405.
Coy, R.C. (1998), Practical Applications of Lubrication Models in Engines, Tribology International, 31
(10):563-571.
Deckman, D. (1995), Chemical aspects of silicon carbide boundary lubrication: mechanisms of sulfur
containing additives, Ph.D. Thesis, Pennsylvania, State University.
Dickinson, J.T., Langford, S.C., and Jensen, L.C. (1993), Recombination on Fractal networks — photon and
electron-emission following fracture of materials, Journal of Materials Research, 8 (11):2921-2932.
Dowson, D. and Higginson, G.R. (1959), A numerical solution to the elastohydrodynamic problem,
J. Mech. Eng. Sci., 1(1):6-15.
Ducker, W.A., Senden, T.J., and Pashley, R.M. (1991), Direct measurement of colloidal forces using an
atomic force microscope, Nature, 353 (6341):239-241.
Fein, R.S. (1960), Transition temperatures with four-ball machine, ASLE Trans., Vol. 3, 34.
Feldman, K., Fritz, M., Hahner, G., Marti, A., and Spencer, N.D. (1998), Surface forces, surface chemistry
and tribology, Tribology International, 31 (1-3):99-105.
Fischer, D.A., Hu, Z.S., and Hsu, S.M. (1997a), Molecular orientation and bonding of monolayer stearic
acid on copper surfaces prepared in air, Tribology Letters, 3 (1):41-46.
Fischer, D.A., Hu, Z.S., and Hsu, S.M. (1997b), Tribochemical and thermochemical reactions of stearic
acid on copper surfaces in air as measured by ultra-soft X-ray absorption spectroscopy, Tribology
Letters, 3 (1):35-40.
Francis, H.A. (1970), Interfacial temperature distribution within a sliding hertzian contact, ASLE Trans.,
Vol. 14, 41.
Furey, M.J. (1964), Surface temperature in sliding contact, ASLE Trans., 7:133-146.
Gates, R.S. and Hsu, S.M. (1995), Silicon nitride boundary lubrication — lubrication mechanism of
alcohols, Tribology Trans., 38 (3):645-653.
Gates, R.S., Hsu, S.M., and Klaus, E.E. (1989), Tribochemical mechanism of alumina with water, Tribology
Trans., 32 (3):357-363.
Gates, R.S., Jewett, K.L., and Hsu, S.M. (1989), A study on the nature of boundary lubricating film —
analytical methods of development, Tribology Trans., 32 (4):423- 430.
Georges, J.M., Tonck, A., Poletti, S., Yamaguchi, E.S., and Ryason, P.R. (1998), Film thickness and
mechanical properties of adsorbed neutral and basic zinc diisobutyl dithiophosphates, Tribology
Trans., 41 (4):543-553.
Gibson, C.T., Watson, G.S., and Myhra, S. (1996), Determination of the spring constants of probes for
force microscopy/spectroscopy, Nanotechnology, 7 (3):259-262.
Godfrey, D. (1962), Chemical changes in steel surfaces during extreme pressure lubrication, ASLE Trans.,
Vol. 5, 57.
Granick, S. (1991a), Motions and relaxations of confined liquids, Science, 253 (5026):1374- 1379.
© 2001 by CRC Press LLC
Granick, S. (1991b), Molecular tribology, MRS Bulletin., 16 (10):33-35.
Granick, S. (1999), Soft matter in a tight spot, Physics Today, 52 (7):26-31.
Greenwood, J.A. and Williamson, J.B.P. (1966), Contact of nominally flat rough surfaces, Proc. R. Soc.
Lond., Vol. A295, 300-319.
Gunderson, R.C. and Hart, A.W. (1962), Synthetic Lubricants, Reinhold, New York.
Gunsel, S., Spikes, H.A., and Aderin, M. (1993), In situ measurement of Zddp films in concentrated
contacts, Tribology Trans., 36 (2):276-282.
Harrison, J.A. and Brenner, D.W. (1994), Simulated tribochemistry — an atomic-scale view of the wear
of diamond, Journal of the American Chemical Society, 116 (23):10399- 10402.
Harrison, J.A., Colton, R.J., White, C.T., and Brenner, D.W. (1993), Effect of atomic-scale surface-roughness
on friction — a molecular-dynamics study of diamond surfaces, Wear, 168 (1-2):127-133.
Harrison, J.A., White, C.T., Colton, R.J., and Brenner, D.W. (1995), Investigation of the atomic-scale friction
and energy-dissipation in diamond using molecular-dynamics, Thin Solid Films, 260 (2):205-211.
Hazel, J.L. and Tsukruk, V.V. (1998), Friction force microscopy measurements: normal and torsional spring
constants for V-shaped cantilevers, Journal of Tribology — Transactions of the ASME, 120 (4):814-819.
Hermance, H.W. and Egan, T.F. (1958), Organic deposits on precious metal contacts, Bell System Tech. J.,
Vol. 27, 739-766.
Hoogerbrugge, P.J. and Koelman Jmva (1992), Simulating microscopic hydrodynamic phenomena with
dissipative particle dynamics, Europhysics Letters, 19 (3):155-160.
Hsu, S.M. (1991), Boundary lubrication of advanced materials, Mrs Bulletin., 16 (10):54- 58.
Hsu, S.M. and Klaus, E.E. (1978), Estimation of molecular junction temperatures in a four-ball wear
contact by chemical reaction rate studies, ASLE Trans., Vol. 21, (3):201.
Hsu, S.M. and Klaus, E.E. (1979), Some chemical effects in boundary lubrication part 1: base oil-metal
interaction, ASLE Trans., Vol. 22, (2):135-145.
Hsu, S.M., Klaus, E.E., and Cheng, H.S. (1988), A mechano-chemical descriptive model for wear under
mixed lubrication conditions, Wear, 128 (3):307-323.
Hsu, S.M., Shen, M.C., Klaus, E.E., Cheng, H.S., and Lacey, P.I. (1994), Mechanochemical model —
reaction temperatures in a concentrated contact, Wear, 175 (1-2):209-218.
Hucknell, D.J. (1974), Selective Oxidation of Hydrocarbons, Academic Press, New York.
Israelachvili, J. (1993), Dynamic properties of liquids at interfaces, Abstracts of Papers of the American
Chemical Society, 206:103-COLL.
Israelachvili, J.N. (1986), Measurement of the viscosity of liquids in very thin-films, Journal of Colloid
and Interface Science, 110 (1):263-271.
Israelachvili, J.N., Mcguiggan, P.M., and Homola, A.M. (1988), Dynamic properties of molecularly thin
liquid-films, Science, 240 (4849):189-191.
Jaeger, J.D. (1942), Moving sources of heat and the temperature at sliding contacts, Proc. R. Soc. London,
N.S.W., Vol. 56, 203.
Johnson, J.C., (1975), Antioxidants: Synthesis and Applications, Noyes Data Corp., Park Ridge, N.J.
Kang, H.J., Zhao, Q., Talke, F.E., Perettie, D.J., Dekoven, B.M., Morgan, T.A., Fischer, D.A., Hsu, S.M.,
and Bhatia, S. (1999), The use of cyclic phosphazene additives to enhance the performance of the
head/disk interface (C), Lubrication Engineering, 55 (3):22-27.
Klaus, E.E., Duda, J.L., and Chao, K.K. (1991), A study of wear chemistry using a micro sample fourball wear test, STLE Tribology Trans., 34 (3):426.
Klaus, E.E., Krishnamachar, V., and Dang, H. (1980), Evaluation of basestock and formulated lubes using
the Penn State micro-oxidation test, Proceedings of the Conference on Measurements and Standards
for Recycled Oil, Vol. 584, 285.
Klaus, E.E., Nagarajan, R., Duda, J.L., and Shah, K.M. (1987), The adsorption of tribochemical reaction
products at solid surfaces, Proceedings of the Inst. Mech. Engr. International Conference on Friction,
Lubrication and Wear, Vol. 1, 379.
© 2001 by CRC Press LLC
Klaus, E.E., Shah, P., and Krishnamachar, V. (1982), Development and use of the micro-oxidation test
with crankcase oils, Proceedings of the Conference on Measurements and Standards for Recycled Oil,
National Bureau of Standards, Gaithersburg, MD.
Klaus, E.E. and Tewksbury, E.J. (1973), Microcorrosion studies with functional fluids, Lubrication Engineer, Vol. 29, 205.
Klaus, E.E., Ugwuzor, D.I., Naidu, S.K., and Duda, J.L. (1985), Lubricant metal interaction under conditions simulating automotive bearing lubrication, Proceedings of the JSLE International Conference,
JSLE, 859.
Kong, Y., Manke, C.W., Madden, W.G., and Schlijper, A.G. (1997), Effect of solvent quality on the
conformation and relaxation of polymers via dissipative particle dynamics, Journal of Chemical
Physics, 107 (2):592-602.
Krim, J. (1996), Friction at the atomic scale, Scientific American, 275 (4):74-80.
Lahijani, J., Lockwood, F.E., and Klaus, E.E. (1981), The influence of metals on sludge formation, ASLE
Transactions, 25(1):25-32.
Lee, S.C. and Cheng, H.S. (1992), On the relation of load to average gap in the contact between surfaces
with longitudinal roughness, Tribology Trans., 35 (3):523-529.
Lenahan, P.M. and Curry, S.E. (1990), First observation of the (29) Si hyperfine spectra of silicon dangling
bond centers in silicon nitride, Appl. Phys. Lett., Vol. 56, (157):207.
Lindsay, N.E., Carter, R.O., Schmitz, P.J., Hack, L.P., Chase, R.E., deVries, J.E., and Willermet, P.A. (1993),
Characterization of films formed at a lubricated cam/tappet contact, Spectrochemica Acta, Vol. 49A,
2057.
Ling, F.F., Klaus, E.E., Fein, R.S., and Editors (1969), Boundary Lubrication: An Appraisal of World
Literature, ASME, New York.
Martin, J.M., Belin, M., and Mansot, J.L. (1986), Friction-induced amorphization with ZDDP — an
EXAFS study, ASLE Trans., Vol. 49, 523.
Mate, C.M. (1995), Force microscopy studies of the molecular origins of friction and lubrication, IBM
Journal of Research and Development, 39 (6):617-627.
McCool, J.L. (1986), Comparison of models for the contact of rough surfaces, Wear, Vol. 107, 37-60.
Meeks, S.W., Weresin, W.E., and Rosen, H.J. (1995), Optical-surface analysis of the head-disk-interface
of thin-film disks, Journal of Tribology — Transactions of the ASME, 117 (1):112-118.
Mizuhara, K. and Hsu, S.M. (1992), Tribochemical reaction of oxygen and water on silicon surfaces, in
Wear Particles, Dowson, D. (Ed.), Elsevier Science Publishers, 323.
Morecroft, D.W. (1971), Reaction of octadecane and decoic acid with clean iron surfaces, Wear, 18:333.
Mori, S. and Imaizumi, Y. (1988), Adsorption of model compounds of lubricant on nascent surfaces of
mild and stainless steels under dynamic conditions, STLE Trans., Vol. 31, 449.
Naidu, S.K., Klaus, E.E., and Duda, J.L. (1984), Evaluation of liquid phase oxidation products for ester
and mineral oil lubricants, Ind. Eng. Chem. Prod. Res. Dev., Vol. 23, 613.
Naidu, S.K., Klaus, E.E., and Duda, J.L. (1986), Kinetic model for high-temperature oxidation of lubricants, Ind. Eng. Chem. Prod. Res. Dev., Vol. 25, 566.
Nakayama, K. and Hashimoto, H. (1991), Triboemission from various materials in atmosphere, Wear,
147 (2):335-343.
O’Connor, J.J. (1968), Standard Handbook of Lubrication Engineering, McGraw Hill, New York.
Perry, M.D. and Harrison, J.A. (1996), Molecular dynamics studies of the frictional properties of hydrocarbon materials, Langmuir, 12 (19):4552-4556.
Quinn, T.F.J., Sullivan, J.L., and Rowson, D.M. (1984), Origins and development of oxidational wear at
low ambient-temperatures, Wear, 94 (2):175-191.
Rabinowicz, E., (1965), Friction and Wear of Materials, John Wiley & Sons: New York.
Ramsey, J.A. (1967), The emission of electrons from aluminium abraded in atmosphere of air, oxygen,
nitrogen, and water vapour, Surface Science, Vol. 8, 313.
© 2001 by CRC Press LLC
Ranney, M.W. (1980) Synthetic Oils and Additives for Lubricants: Advances since 1977, Noyes Data Corp.,
Park Ridge, N.J.
Reynolds, O. (1886), On the theory of lubrication and its application to Mr Beauchamp Tower’s experiments, including an experimental determination of the viscosity of olive oil, Phil. Trans. R. Soc.,
117:157-234.
Rosenblum, F., Braunlich P., and Himmel, L. (1977), Spontaneous emission of charged particles and
photons during tensile deformation of oxide-covered metals under ultrahigh vacuum conditions,
J. Appl. Phys., Vol. 48, 5262.
Satriana, M.J. (1982) Synthetic Oils and Additives: Advances Since 1979, Noyes Data Corp., Park Ridge, N.J.
Schneir, J., Mcwaid, T.H., Alexander, J., and Wilfley, B.P. (1994), Design of an atomic-force microscope with
interferometric position control, Journal of Vacuum Science & Technology B, 12 (6):3561-3566.
Sheasby, J.S., Caughlin, T.A., and Habeeb, J.J. (1991), Observation of the antiwear activity of zinc
dialkyldithiophosphate additives, Wear, 150 (1-2):247-257.
Shubkin, R.L. (1993), Synthetic Lubricants and High Performance Functional Fluids, Marcel Dekker, New
York.
Simoi, C., Hrianca, I., and Cracium, P. (1968), Exoemission of electrons without photo-stimulation, Phys.
Status Solidi, Vol. 29, 761.
Smalheer, C.V. and Smith, R.K. Lubricant Additives, Lezius-Hiles, Cleveland, OH.
Spikes, H.A. (1996), Direct observation of boundary layers, Langmuir, 12 (19):4567-4573.
Stuart, S.J. and Harrison, J.A. (1999), Molecular dynamics study of lubrication by small hydrocarbons,
Abstracts of Papers of the American Chemical Society, 217:U589.
Sugimura, J. and Kimura, Y. (1984), Characterization of topographical changes during lubricated wear,
Wear, Vol. 98, 101.
Tomizawa, H. and Fischer, T. (1986), Friction and wear of silicon nitride and silicon carbide in water:
hydrodynamic lubrication at low sliding speed obtained by tribochemical wear, ASLE Trans., Vol. 30, 41.
Tonck, A., Kapsa, Ph., and Sabot, J. (1986), Mechanical behavior of tribochemical films under a cyclic
tangential load in a ball flat contact, J. Tribology, Vol. 108, 117.
Wang, F.X., Lacey, P., Gates, R.S., and Hsu, S.M. (1991), A study of the relative surface conformity between
2 surfaces in sliding contact, Journal of Tribology — Transactions of the ASME, 113 (4):755-761.
Warren, O.L., Graham, J.F., Norton, P.R., Houston, J.E., and Michalske, T.A. (1998), Nanomechanical
properties of films derived from zinc dialkyldithiophosphate, Tribology Letters, 4:189-198.
Yang, Y.Y., Torrance, A.A., and Oxley, P.L.B. (1996), Modelling mechanical wear processes in metallic
sliding friction, Journal of Physics D — Applied Physics, 29 (3):600-608.
Yates, J.T., Dohnalek, Z., Yang, W., and Choyke, W.J. (1996), Defect site effects on chlorine chemistry on
silicon, Abstracts of Papers of the American Chemical Society, 211:18-PHYS.
Ying, T.N. (1994), Wear mechanisms for ductile and brittle materials in a microcontact, Ph.D. Thesis,
Department of Materials and Nuclear Engineering, University of Maryland, College Park, MD.
For Further Information
The reader is directed to the following important resources for further information on the topics discussed
in this chapter on boundary lubrication.
Boundary Lubrication: An Appraisal of World Literature, F.F. Ling, E.E. Klaus, and R.S. Fein, Editors,
ASME, NY [1969].
CRC Handbook of Lubrication, Theory and Practice of Tribology, Volume II: Theory and Design, Richard
Booser, Editor, CRC Press, Boca Raton, FL [1984].
Limits of Lubrication, selected papers from the Limits of Lubrication Conference in Williamsburg, VA,
April 14-18, 1996. Tribology Letters 3(1) [1997].
Tribochemistry, Heinke, G., Hanser Press, Munich, Germany [1984].
© 2001 by CRC Press LLC
Defining Terms
Additive: Chemical compounds dissolved or dispersed in a base fluid to impart or extend desirable
properties to the lubricant.
Basestocks: Main fluid component of a lubricant. Can be derived from naturally occurring, or synthetic
sources.
Boundary Lubrication: Lubrication regime involving contact between asperities in relative motion in
which chemical reactions dominate the mechanism of reduction of friction and wear.
Dangling Bonds: Highly reactive sites on atoms resulting from inability of the atom to form proper bonds
with nearest neighbor atoms. Can occur when bonds in a structure are broken by mechanical forces.
Friction Polymer: High-molecular-weight product formed in some contacts under boundary lubricating conditions.
Kurtosis: Surface roughness parameter describing the narrowness of the peak height distribution.
Nascent Surfaces: Fresh surfaces formed during rubbing. Highly reactive.
Organometallic: Chemical compounds consisting of metal atoms bonded to organic molecules.
Ra: Surface roughness parameter describing the average surface roughness, also known as centerline
average roughness.
RMS: Surface roughness parameter describing the root mean square surface roughness.
Skewness: Surface roughness parameter describing the asymmetry of the peak height distribution of
a surface.
Tribochemistry: Chemical reactions that are induced by rubbing, usually referring to reactions that
are not observed by purely thermal means.
© 2001 by CRC Press LLC
Fly UP