Chapter 11 FRICTION AND WEAR

It is well appreciated that friction and wear are interrelated subjects simply because friction is involved in wear mechanisms. Indeed, both friction and wear can be studied and measured in the same experiment and this is done in, for example, the investigation of bearings and sliding joints. Friction plays its part in rubber wear mechanisms and needs to be considered when these mechanisms are being studied. Furthermore, with a major rubber product, tyres, friction (resistance to slip) and wear are two of the most important performance parameters. However, when it comes to laboratory measurements on rubber the two tests are generally considered as separate subjects and historically have not been accredited equal status. Whereas friction tests have rarely been standardised and are carried out in relatively few laboratories, dozens of wear or abrasion tests have been developed, quite a few have been standardized and abrasion apparatus is relatively common in rubber laboratories. It can be argued that this situation is not unreasonable, as far more than simply friction is involved in rubber wear processes and wear is probably seen to be directly a problem in more applications of rubber than is friction. On the other hand, the role of friction in slipping accidents and in sports activities has assumed huge importance over the last couple of decades. This has resulted in a large amount of testing and test method development which is relevant to rubbers in flooring and footwear.

1.

FRICTION

Friction of rubber is a very complicated phenomenon but is generally thought of as being composed of two parts, adhesive friction and hysteresis

220

Physical testing of rubber

friction. The adhesive friction arises from the repeated formation and breaking of molecular bonds between the surface during sliding, whilst the hysteresis friction is a result of the dynamic deformation of the rubber with energy being lost. From the testing point of view, the main significance of the mechanisms of friction is that we must expect the level of friction to be dependent on both the external factors, such as the roughness of the surface, and the internal factors associated with the rubber's viscoelastic properties.

1.1

Factors affecting friction

It would appear that no account of friction is complete without first stating Leonardo da Vinci's (or Amonton's) laws and Coulomb's law of friction and pointing out that, in general, polymers do not obey them. The laws are: (a) The frictional force opposing motion is proportional to the normal force, the constant of proportionality being the coefficient of friction (see Figure 11.1), i.e.:

where F = frictional force, \x = coefficient of friction and N = normal force.

F = |aN

Figure 11-1. Coefficient of friction

(b) The coefficient of friction is independent of the apparent area of contact. (c) The coefficient of friction is independent of the velocity between the two surfaces provided that the velocity is not zero. In practice, rubbers do not normally obey these rules and the coefficient is a 'variable constant', its value depending on the real contact area, normal load, velocity and other factors. In fact, friction is sensitive to just about

Friction and wear

221

anything, including breathing on the test piece, and any single-point measurement is of very limited use. The factors influencing friction of polymers have been discussed by, for example, James ^ who gives a useful bibliography, and Ohhara^ who outlines the theory of rubber friction. Only the factors of most importance as regards the test method will be considered briefly here. The apparent area of contact between two surfaces is much larger than the actual area over which they touch, even if the surfaces appear smooth. The frictional force is proportional to the real contact area, so anything that changes the real contact area will change the force measured. When rubber is brought into contact with another surface it deforms elastically and the real area of contact will increase with increasing normal load and, hence, the coefficient of friction will decrease with increasing normal load. It is also apparent that the real contact area is dependent on the surface geometry of the test piece. It is, hence, desirable to measure the friction of rubber over the range of normal forces of interest and to test with the surface geometry to be used in service, which may mean using the product or part of it as the test piece. An interesting aspect of contact pressure is that it is not uniform over the surface of a thick rubber block, but increases towards the edge. The effect of this on friction has been discussed by Nakajima and Takahashi^. The distinction is sometimes made between static and dynamic friction, implying that there is one level of the coefficient of friction just at the point when movement between the surfaces starts and another level when the surfaces are steadily separating. There can of course be no measure of friction without movement so that 'static' friction is actually friction at an extremely low velocity and thereafter the coefficient of friction of rubbers may vary markedly with velocity. Hence, it is necessary to measure friction over the range of velocities of interest. Friction is also dependent on temperature, which can lead to inaccuracies at high velocities because of heat build-up at the contacting surfaces. During a friction test, a condition known as 'slip-stick' sometimes occurs in which the relative velocity and the coefficient of friction between the two surfaces both oscillate about a mean value. The essential condition for slipstick to occur is that in the velocity region being considered the coefficient of friction falls with increasing velocity. An over-simplistic description of the phenomenon is that for a short time the surfaces stick together and the force builds up as in a spring until it exceeds the 'static' coefficient of friction when movement occurs and the friction falls to a lower kinetic value and the spring releases until sticking again takes place. Slip-stick can in fact occur at high velocities, when the frequency of vibration can be high enough to cause audible squeals. The amplitude and frequency of the slip-stick vibrations

222

Physical testing of rubber

depend on the rigidity and damping of the testing system as well as on the properties of the surfaces. To minimise slip-stick it is necessary to construct the test apparatus, particularly the drive and force measuring elements, to be as stiff as possible. If slip-stick can occur in service, its presence can be more important, or rather troublesome, than the actual mean level of friction. It is fairly obvious that other factors such as lubricants, wear debris, ageing of the surfaces and humidity can also affect friction and, once again, test conditions must be chosen that resemble those found in service.

1.2

Methods of measuring friction

The essential requirements for a friction test are two contacting surfaces, a means of creating relative motion between them and a system to indicate the frictional force. A number of different arrangements are then possible, a selection being shown in Figure 11.2.
N \
(a)

ZZyA
/ /

'

// V

i"
(c) ,/V

(b)

'm^mmmmmmmm

\N^wcoze

Figure 11-2. Arrangements for friction tests, (a) Linear track; (b) rotating shaft; (c) towed sled; (d) pin and rotating place (e) inclined plane. N = normal force, V = direction of motion, W = weight of test piece

Friction and wear

223

An example of a driven sled apparatus has been described by James"*, Griffin^ devised a test for small cylindrical plastic test pieces and Mustafa and Udrea^ used a rotating steel disc and a stationary plastic test piece. A ball and peg machine is described by Bailey and Cameron^, a hemispherically ended pin and flat plate device by West and Senior^, a steel pin running on a plastic ring by Jost^ and a combined inclined and horizontal plane by Wilson and Mahoney^^. A computerised plane on plane reciprocating apparatus was developed by Benabdallah^^ and a ring on disc was used by Bielinski et al^^. Roberts and Alhston-Greiner^^ developed a machine with ball or cylinder on flat geometry whilst Bandel and Di Bernardo^"^ aimed at simulating tyres with a rotating wheel on a rotating flat arrangement, but this list is by no means exhaustive. The towed sled is the most common form of test and many such apparatus have been devised. Whilst simple in principle, there are practical problems in that the sled will tend to tilt if it is not towed on the plane where the surfaces meet, and if the means of applying the force is a wire or cord, the lack of stiffness can cause slip-stick. An attractive basis for a good quality friction test is a universal tensile testing machine because it can be suitably stiff, gives a very wide range of speeds and has a precise force measuring system. The only difficulty is that such machines operate in a vertical plane and, if the normal load on the test piece is applied by a weight acting vertically under gravity, the linkage to measure the frictional force must turn through 90, and in doing so there is a danger of introducing friction at a pulley and decreasing the stiffness of the system. Appreciating these difficulties of precisely measuring the frictional properties of polymers, a novel apparatus operating vertically in a tensile machine was developed at Rapra and has been described by James and Newell^^. The advantages of the apparatus (Figure 11.3) are that, in conjunction with a suitable tensile tester, a very stiff system results with very accurate measurement of small forces and a good range of velocity. The construction used also overcomes the inherent problem in sled devices of accurately aligning the force direction with the plane of the sliding surfaces. It can be operated in an environmental chamber, so giving a wide range of temperatures and the test piece geometry can be readily changed, including tests on products or parts of products. Using this apparatus, James and Mohsen^^ have illustrated the importance of specifying sample preparation and test conditions for rubber. Product areas where friction of rubber is particularly important are roads and floor surfaces where it is convenient if measurements can be made insitu. Consequently, a considerable number of portable devices have also been developed.

224
Load cell

Physical testing of rubber

Pivot

Figure US. Rapra friction apparatus

The friction of road surfaces is often measured with a skid tester developed by the Road Research Laboratory'"^ and this has also been widely used on other surfaces and floors, including artificial sports surfaces. It is a pendulum device, the movement of which is arrested by the foot of the pendulum skidding on the surface to be measured. The 'skid resistance' indicated can be approximately related to coefficient of friction by: Skid resistance-

330// 3 + //

Friction and wear

225

A number of other instruments have also been developed for this purpose, including the Tortus apparatus^^, which is essentially a selfpropelled, four wheeled trolley with a sliding foot, and various portable towed sleds. Other devices developed for sports surfaces are the Leroux pendulum, which is similar in principle to the skid tester, a traction apparatus that works in torsion and a sliding resistance tester comprised of a large trolley given momentum by a ramp. The use of these methods in artificial sports surface testing has been discussed by Brown^^. Tests for flooring and footwear are not always portable, some utilise a whole shoe with linear motion such, as the Satra slip test^^'^\ and many workers favour walking on a full sized ramp^^. The reasoning behind test methods for slip resistance and descriptions of several apparatus, together with a large bibliography, have been given by James^^. James pointed out a long time ago the importance of kinetic friction in shoe/floor interaction^"^ but strut devices measuring "static" friction have been popular in the USA^^' ^^. Summaries of European investigations have been given^'^' ^^ but there is no doubt that the subject of pedestrian friction is extremely complex and there is no general agreement as to the best measurement methods. The Hughes gauge, a really novel but simple apparatus which would let us all do our own slip measurements, relies solely on the geometry of the slider^^. Because of the sensitivity of friction to so many variables, it is often desirable, if not essential, to test the actual product in a prototype test rig. Certain of the slip resistance tests have headed in that direction by using complete shoes as the slider. Similarly, instrumented systems are used to investigate the grip of tyres under real road conditions. Generally, a product test rig will be the best approach for items such as bottle closures and bearings. Reviews of methods and equipment for friction and wear of rubbers have been given by Mitsuhashi^^ and Sviridyonok and Kirpichenko"^^ and a number of comparisons between machines have been reported^^"^^. Correlation between different methods and with service can be relatively poor, which is perhaps not surprising considering the variety of geometries and test conditions in use and the complexity of the interaction between surfaces in such actions as walking.

1.3

Standard methods

An ISO standard for the frictional properties of rubber was published in 1999^^ and is probably one of the most comprehensive of friction standards. It is based on plane on plane geometry on the basis that rubber samples are most readily available in sheet form and for many applications measurement between two planar surfaces most nearly approaches service behaviour.

226

Physical testing of rubber

However, it recognises that with this basic geometry involving linear motion it is often possible to substitute other shapes for one of the sliding members, for example a product such as a windscreen wiper blade. A problem with plane on plane geometry is that if the friction plane does not contain both the line of action of the load cell and the line of the towed force then there will be a tilting moment which introduces errors. For this reason, the standard notes the preference of some workers to use a ball on plane geometry^^. Whilst this overcomes one problem, frictional force is then not proportional to normal load and the contact area must be estimated from knowledge of the modulus of the rubber. Lining up of the friction plane with load cell and towing force is achieved with plane on plane using an apparatus such as described by James and Newall^^. The standard is not one apparatus specific but requires tight control of the important parameters and gives considerable guidance both in the text and in annexes on factors to be considered in making friction measurements. It allows for testing at a number of velocities and normal loads and procedures are given for preparing the sliding surfaces. Three procedures for determining dynamic friction are given, the initial friction, friction after repeated movement between the surfaces and friction in the presence of lubricants or contaminants. The presentation of results gives considerable detail on the interpretation of the friction traces, including dealing with shp stick. The ISO standard was developed from a British standard and the two are now identical, the British standard being numbered both BS ISO 15113 and BS 903 Part A 61^^ ASTM does not have a method for determination of rubber friction. An inclined plane method is specified for coated fabrics in BS3424"^^. The test piece is attached to a sled resting on an inclined plane which is covered with the other surface to be tested. The inclination of the plane is varied until sliding takes place. This is probably the simplest form of friction test but not an accurate one and, of course, cannot operate at any given velocity. BS 3424 also has a simple towed sled method which is very briefly described such that there is too little control of apparatus parameters for it to be adequate for more general or accurate work. The drive linkage is not specified in detail and some apparatus in use would lack stiffness. A better text for a towed sled is given in the international, British and European standard for plastics film"^^ which recognises the importance of parameters such as stiffness of the drive. When evaluating the friction of surfaces such as roads or sports floors, it is necessary to standardize in some way the slider. Not surprisingly, leather has been used but it is can be variable. An early development was the rubber

Friction and wear

22 7

used in the skid tester and since then there has been detailed investigation leading to a rubber particularly suited to pedestrian friction"^^' ^'^. The objective was to have a material that gives the best correlation with service experience together with good reproducibility between laboratories. Any rubber standardised for this purpose must be produced to a very precise specification and the material now very widely used is controlled by its dynamic properties at a number of temperatures.

2.

WEAR

The terms wear and abrasion are used so loosely that confusion sometimes results. Wear is a very general term covering the loss of material by virtually any means. The dictionary says that abrasion is the wearing away by means of friction, although in everyday life we think of it as the rasping action of a rough surface. As wear usually occurs by the rubbing together of two surfaces, abrasion is often used as a general term to mean wear. The mechanisms by which wear occurs when a rubber is in moving contact with any material are somewhat complex, principally involving cutting of the rubber and fatiguing of the rubber. Nevertheless, we call the tests to measure this wear abrasion tests.

2.1

Wear mechanisms

An account of the mechanisms of both friction and wear has been given by Lancaster"^^ and one section of a collection of translations of papers"^"^ describing extensive Russian work on abrasion is devoted to mechanisms. Further discussion of wear mechanisms is given, for example, by Gent and Pulford"^^ and Schallamach'*^"'*^. Efforts have been made to find a relationship between wear and fundamental properties of rubber"^^ and a review has been presented^^ which attempts to clarify what is understood about the abrasion process and what is still unclear. It is possible to categorise wear mechanisms of rubber in various ways and one convenient system is to differentiate between three main factors: (a) Abrasive wear, which is caused by hard asperities cutting the rubber. (b) Fatigue wear, which is caused by particles of rubber being detached as a result of dynamic stressing on a localised scale. (c) Adhesive wear, which is the transfer of rubber to another surface as a result of adhesive forces between the two surfaces.

228

Physical testing of rubber

Wear by roll formation is where there is progressive tearing of a layer of rubber which forms a roll. The result is a characteristic abrasion pattern of ridges and grooves at right angles to the direction of movement. The term erosive wear can be applied to the action of particles conveyed in a liquid stream and there can also be corrosive wear due to direct chemical attack of the surface. From these definitions, it can be seen that the more specific meaning of abrasion is wear by the cutting action of hard asperities. The common practice in the rubber industry of using abrasion as a general term for wear probably results from the fact that most wear tests for rubbers use the action of sharp asperities, for example abrasive paper, to produce wear. The wear or abrasion of rubber caused by hard asperities is not just simply cutting but involves both plastic and elastic deformation of the rubber^ ^ However, abrasive wear requires the abradant to have hard, sharp cutting edges and high friction, while fatigue wear occurs with smooth or rough but blunt surfaces and does not need high friction. Adhesive wear is not so important for rubbers but can occur on smooth surfaces. Roll formation requires high friction and relatively low tear strength. Fatigue processes produce much less rapid wear than cutting processes. Generally, in any wear process more than one mechanism is involved although one mechanism may predominate. The mechanism, and hence the rate of wear, can change with change of conditions such as contact pressure, speed and temperature. The most important consideration in practice is that the wear process will be complex and critically dependent on the service conditions. It is, therefore, necessary that any laboratory test must essentially reproduce the service conditions if good correlation is to be obtained. Even a comparison between two rubbers may be invalid if the predominant wear process in the test is different from that in service. It is failure fully to appreciate this which has led to the conclusion that all laboratory abrasion tests are useless except for quality control. It follows, that there cannot be a universal standard abrasion test for rubber and the test method and conditions have to be chosen to suit the end application. In some applications, for example tyres, the range of conditions encountered is so complex that they cannot be matched by a single laboratory test. However, for many products meaningful results can be obtained by careful modification of standard abrasion tests, but great care has to be taken if the test is intended to provide a significant degree of acceleration.

Friction and wear

229

2.2

Types of abrasion test

A large number of different abrasion apparatus have been used for testing rubbers and an even larger number of permutations of the various factors would be possible. The first division of test types can be to distinguish between those using a loose abradant and those using a 'solid' abradant. A loose abrasive powder can be used to impinge on the rubber rather in the manner of a shot blasting machine, or tumbled with the rubber test pieces in a rotating drum. These are logical ways to simulate the action of sand or similar abradants impinging on the rubber in service, as may be the case with conveyor belts or tank linings, but this type of test is not very common. 'Solid' abradants could consist of almost anything but the most common are abrasive wheels (vitreous or resilient), abrasive papers or cloth, and metal 'knives'. The possible geometries by which the test piece and a solid abradant can be rubbed together are legion and it is not sensible to make any general classification. Some well-known configurations are shown in Figure. 11.4. In type (a) the test piece is reciprocated linearly against a sheet of abradant, but alternatively a strip of abradant could be moved past a stationary test piece. A further variation is to have the abradant as a rotating disc with the test piece held against its side (b). Both the abradant and the test piece can be in the form of wheels, type (c), with either being the driven member. In type (d) the abrasive wheel is driven by a rotating flat test piece, and in type (e) both the test piece and abradant are rotated in opposite directions. Type (f) the test piece is held against a rotating drum with the abradant on its surface. The test piece can also be made to traverse along the length of the drum. A loose abradant can also be used between the two sliding surfaces in what could be considered as a hybrid of loose and solid abradant tests. This situation occurs in practice through contamination and as a result of the generation of wear debris from a 'solid' abradant. A car tyre is an example of the situation where there is a combination of abrasion against a solid rough abradant, the road, together with a free flowing abradant in the form of grit particles.

2.3

Abradants

Abrasive wheels are characterized by the nature of the abrading particles (their size and sharpness), the structure of the wheel and the manner in which the abrasive is bonded (resilient of vitreous). Clearly, a wide range of abrasive properties is possible. Wheels are probably the most convenient abradant because of their low cost, mechanical stability and the fact that a consistent surface can be maintained by simple refacing.

230

Physical testing of rubber

S}-

Figure 11-4. Arrangements for abrasion tests

Abrasive papers and cloths are cheap and easy to use but their cutting power deteriorates rather quickly. They are also characterized by the nature of the abrasive particles and their size and sharpness. Plain textiles of defined quality have also been used for mild abrasion.

Friction and wear

231

Loose abradants are usually particles of the same types of material that are used to form abrasive wheels or paper, and are characterized in the same way. Metal "knives" can have various geometries, including a mesh and a raised pattern on a wheel. The important characteristic is the sharpness of the edges in contact with the rubber test piece, and this can be difficult to accurately maintain. Plane smooth surfaces are usually metal and are characterized by the material and the surface roughness. The choice of abradant should be made primarily to give the best correlation with service, and the usual abrasive wheels and papers really only relate to situations where cutting abrasion predominates. Materials such as textiles and smooth metal plates may be more appropriate for other applications. In practice, the abradant is often chosen largely for reasons of convenience and surfaces such as plain steel have the disadvantage of abrading slowly and, if the conditions are accelerated, give rise to excessive heat build-up. Consequently, abrasive wheels and papers are used in situations where they are inappropriate for assessment of service performance.

2,4

Test conditions

Abrasion occurs when the rubber slips relative to the abradant and the amount of slip is a critical factor in determining the rate of wear. In type (a) of Figure 11.4 the slip is 100% because the rate of slipping is equal to the rate of movement of the test piece (or abradant). In contrast, with the type (c) arrangement a range of levels can be used by varying the skew angle between the two wheels, or in type (d) by varying the distance of the wheel from the centre line of the test piece. Typically, the rate of abrasion with a type (c) apparatus is proportional to something between the square and cube of the slip angle. An important difference between apparatus of type (a), (b), (e) or (f) and (c) or (d) is that in the former case the test piece is continuously and totally in contact with the abradant and there is no chance for the very considerable heat generated at the contact surface to be dissipated. The actual rate of slip will influence the rate of wear because, as the speed is increased, heat buildup will rise. Temperature rise during test is one of the important factors in obtaining correlation between laboratory and service. The contact pressure between the test piece and abradant is another critical factor in determining wear rate. Under some conditions, wear rate is more or less proportional to pressure but, if with changing pressure the abrasion mechanism changes, perhaps because of a large rise in temperature.

232

Physical testing of rubber

then the wear rate may change quite drastically. Again, this is a critical factor in obtaining correlation with service. Rather than consider contact pressure and degree of slip separately, it has been proposed^^ that the power consumed in dragging the rubber over the abradant should be used as a measure of the severity of an abrasion test. The power used will depend on the friction between the surfaces and will determine the rate of heat build-up. Although temperature has a large effect on wear rate, it is extremely difficult to control the temperature during test, but it is clearly the temperature of the contacting surfaces which is of importance rather than the ambient temperature. The rate of wear will quite naturally be affected by any change in the nature of the contacting surfaces. Apart from the abradant changing because of its own wear, there can be effects from lubricants, wear debris between the surfaces and clogging of the abradant. Not many commonly used apparatus are suitable for testing in the presence of a liquid lubricant, but it is common practice to remove wear debris by continuously brushing the test piece or by the use of air jets, in which case care must be taken to ensure that the air supply is not contaminated with oil or water from the compressor. Clogging or smearing of the abradant is a common problem with abrasive wheels and papers and its occurrence will invalidate the test. It is normally caused by a high temperature at the contact surfaces and, although the problem can sometimes be reduced by introducing a powder between the surfaces, it should be treated as an indication that the test conditions are not suitable. If high temperatures are to be realised in service, a test method in which new abradant is continually used should be chosen. It may be noted that a practical example of a powder influencing abrasion is a car tyre running on a dusty surface. If the abrasion is unidirectional, abrasion patterns will develop which can markedly affect abrasion loss. It should again be emphasised that, if correlation between laboratory tests and service is to be obtained, the test conditions must be chosen extremely carefully to match those found in the product application.

2.5

Expression of results and standard rubbers

In standard abrasion tests, it is usually weight loss which is the parameter measured, although in certain cases the change in test piece thickness is more convenient. Because it is the amount of material lost which matters, it is usual to convert the weight loss to volume loss by dividing by the density. The volume loss can be expressed as the loss per unit distance travelled over the abradant, per 1000 revolutions of the apparatus, or whatever. A less usual practice is to express the result as loss per unit energy consumed in

Friction and wear

233

causing abrasion which is sometimes referred to as abradability and can be linked to the sHding speed through the Wilhams Landel Ferry (WLF) relation. The volume loss may also be calculated per unit surface area to give a specific wear rate. Whatever the loss is related to, it must be remembered that the rate of wear may not be constant because of inhomogeneity of the test piece and gradual change in the nature of the abradant. The experiment should be designed to minimise the latter effect by using standard rubbers, refacing the abradant and running repeat test pieces of a series of materials in reverse order. To investigate test piece or abrasive effects, wear rate can be plotted against number of cycles or the distance travelled. Largely because of the critical dependence of the wear rate on the test conditions, and particularly because of the difficulty in maintaining a precisely reproducible abradant, it is common practice to refer all abrasion results to the results obtained at the same time on a 'standard' rubber. This is an eminently sensible practice as it goes a long way towards eliminating variability due to differences between nominally identical machines and abradants. There is only one drawback, the difficulty of producing an accurately reproducible standard rubber. This produces something of a chicken and egg situation where it is difficult to decide whether it is the abradant or the rubber which has changed. There does not appear to be evidence to demonstrate the variability of either abrasion standards or abradants but within one laboratory the coefficient of variation of abrasion results using different batches of a standard rubber would probably be not much less than 8%, and the between laboratory variation could clearly be very much greater. Some abradants will certainly be more variable than this but other materials can be reproduced with better precision. Although it is a fact that standard rubbers are themselves variable, they are of very considerable value, particularly when reference is made only to standards from one batch and where they are used to monitor the change with time of one sample of abradant or to compare a number of abradants. Standard rubbers are given in several of the test method standards. Some are purely for normalizing the abrasive and have no particular service relevance. Others, such as those based on a tyre tread compound or a shoe sole type material, have the advantage of a practical significance as well as a normalizing role. There is no reason at all why, for a particular investigation, an in-house standard representative of the type of material being evaluated should not be used. Devotees of the use of standard rubbers then finally express the result as an abrasion resistance index defined by:

234

Physical testing of rubber V Abrasion resistance index = ^ x 100%

where Vs = volume loss of standard rubber and Vr = volume loss of rubber under test. Abrasion resistance is the reciprocal of volume loss. If the volume loss or abrasion resistance only is quoted, it is desirable to have some certification of the abradant used. This is naturally supplied to some extent by specifying a particular grade and source of supply but leaves open to question the variability of that source of supply. Some workers prefer to use a standard rubber to test the abradant and to calculate a relative volume loss: V Relative volume loss = -^ xV.

where Vr = volume loss of rubber under test, Vd = the defined volume loss of the standard rubber and Vs = the measured volume loss of the standard rubber. Whichever approach you take, the result is still dependent on the variability of the standard rubber and, arguably, it could be better to rely on the reproducible manufacture of, for example, an abrasive wheel. It would not seem beyond the bounds of ingenuity to find a standard material which is inherently more reproducible than rubber! It could then be used either to certify the abradant or to use in the calculation of abrasion index. If abrasion loss is measured as a function of test parameters such as speed, temperature, degree of slip, contact pressure etc, it may be possible to combine the results in some way to produce a composite measure of abrasion resistance. Obtaining data as a function of test parameters is impossible, or at least very tedious, with most apparatus, but can be achieved automatically with the LAT system (see Section 2.6).

2.6

Test apparatus

There is an International standard for abrasion in the form of a guide^^. This covers wear mechanisms, types of abrasion test, test conditions, procedures and expression of results, as discussed above, as well as giving summaries of thirteen particular abrasion apparatus. The content of the guidance document was in fact developed from an earlier version of this chapter. For many years, the only abrasion test method standardized internationally was the apparatus commonly known as the DIN abrader

Friction and wear

235

because the test was based on the German method. This was standardized because of its very widespread use for quahty control but the standardization of other methods was resisted to avoid a potential proliferation of methods. One objective of having the guidance standard was to allow various instruments to be referenced in an international standard but to avoid multiple standards. As will be seen below, this attempt to restrict the number of standards seems to have failed. The DIN method is given in ISO 4649^"^. The principle of the machine is illustrated in Figure 11.5; a disc test piece in a suitable holder is traversed across a rotating drum covered with a sheet of the abradant, which is why it is also called the rotary drum abrader. In this way, there is a relatively large area of abradant, each part of which is passed over in turn by the test piece, so that wear of the abradant is uniform and relatively slow. In the standard method there is no provision for changing conditions from those specified, other than a lower force for soft rubbers, but it would be possible to use other abradant cloths or papers and to vary the force on the test piece. The degree of slip is 100% and it would be inconvenient to test in the presence of a lubricant. Although not versatile, the method is very convenient and rapid and well suited to quality control.

ABRADANT

TEST PIECE

Figure 11-5. Principle of DIN abrader

The details of procedure and expression of results are something of a compromise, being a compilation of the German approach and the British approach. Two procedures are specified, using a rotating or non-rotating test piece respectively. In principle, the abrasion should be more uniform if the test piece is rotated during test. The standard abradant is specified in terms

236

Physical testing of rubber

of weight loss of a standard rubber using a non-rotating test piece and has to be run in against a steel test piece before use. Results can either be expressed as a relative volume loss with the abradant normalized relative to a standard rubber or as an abrasion index relative to a standard rubber. The former is the original German approach and the latter is that approach favoured by, for example, British standards. Two standard rubbers are specified which originate from the German and British standards. This is a fine example of the chicken and egg problem of standard rubber and abradant. To understand the situation it must be appreciated that an accurately specified standard abradant has been available in Germany for many years, but the rubber used to check it has a formulation of no relevance to real products and is very difficult to reproduce in laboratories other than that of its origin. The British standard rubber is based on a tire tread and, hence, has practical relevance. Either way of expressing results can be used with either rotating or nonrotating test pieces but, normally, only the German standard rubber would be used with relative volume loss. In previous versions, only the abrasion index approach had the option of using a rotating test piece. There is an identical British standard numbered as BS ISO 4649 which has partially superceded BS 903 Part A9 to update the DIN method. This leaves BS 903 A9^^ with procedures for the DuPont, Akron and Taber apparatus. The DuPont apparatus uses a disc of abrasive paper which rotates whilst a pair of moulded test pieces are continuously pressed against it (type (b) of Figure 11.4) either with a constant force or with a force adjusted to give a constant torque on the arm holding the test pieces. It is a relatively simple apparatus using an easily replaced abradant but has several disadvantages. The abrasive paper is prone to smearing with soft materials due to heat build-up, few parameters can be varied and the irregular-shaped test piece has to be specially prepared. The Akron machine is of the form (c) in Figure 11.4. The test piece is a moulded wheel driven at constant speed and held against the abrasive wheel by a constant force. Its main advantage is that by varying the angle of the test piece relative to the wheel the degree of slip can be varied and, hence, its effect studied. Any point on the test piece is not continuously in contact with the abradant so that heat build-up is less troublesome than with the DuPont. It is not particularly convenient to change the abradant and the test piece must be specially prepared. The standard allows for the addition of a loose powder between the test piece and the wheel to prevent clogging of the wheel with soft rubbers. For both the DuPont and Akron procedures, the standard specifies a trial run to establish the level of abrasion rate and a running in period before the

Friction and wear

23 7

actual test runs. The running in period is to bed in the test piece and abradant, and for the DuPont also serves to condition a new piece of abrasive paper. The 'rotary-platform, double-head' or Taber abrader, unlike those mentioned above, was not developed by the rubber industry but was intended for very general use. It is of the form (d) in Figure 11.4 but uses a pair of abrasive wheels. Although the degree of slip cannot be varied, the Taber is in other ways a very versatile apparatus. It uses a simple flat disc as the test piece which could, if necessary, be fabricated from more than one piece. The force on the test piece and the nature of the abradant are very readily varied and tests can be carried out in the presence of liquid or powder lubricants. When using the usual type of abrasive wheel, a refacing procedure is carried out before each material tested. Four standard rubbers are specified, two of which are the same as those given in ISO 4649. The other two are a high abrasion resistant SBR compound and a low abrasion resistant material which may be preferred for comparison with high and low abrasion resistant test materials respectively. For all three methods, the calculation of abrasion resistance index is specified. ASTM has the DIN abrader specified as D5963^^. It has the rotating and non-rotating methods and both standard rubbers but does not cater for the lower force for soft rubbers. The details are said to be equivalent to the 1985 version of ISO 4649 and there is no indication of why it has not been updated. The Pico abrader is specified in D2228^^. This device uses a pair of tungsten carbide knives which rub the test piece whilst it rotates on a turntable. The direction of rotation is reversed at intervals throughout a test and a dusting powder is fed to the test piece surface, which doubtless helps to avoid stickiness. The apparatus is calibrated by the use of no less than five standard rubbers and the result also expressed as an abrasion index. Force on the test piece and speed of rotation can be varied and, presumably, different abradant geometries could be used, although the distinctive feature of the Pico is the use of blunt metal knives in the presence of a powder. In addition, the rotary-platform, double-head abrader is specified for coated fabrics^^ and the NBS abrader for shoe soles and heels^^. The NBS abrader uses rotating drums with abrasive paper wrapped around them onto which the test pieces are pressed by means of levers and weights, but there is no provision for traversing the test piece across the abradant as in the DIN machine. No attempt will be made to give a comprehensive survey of abrasion testers as so many designs have been tried and even some of the standardised types of apparatus are not very commonly used. A review was given by

238

Physical testing of rubber

Buist in 1950^" but many machines have appeared since them. Twenty one methods are listed in a study of wear of flooring materials^^ and several machines developed in the USSR are described in Section 3 of reference 44. The Lamboume abrader is essentially the Dunlop apparatus which used to be specified in BS 903, and which uses a wheel test piece with the degree of slip being controlled by an eddy current brake. An improved Lambourne machine is now being proposed for standardization in ISO TC 45. This instrument has both the test piece and the abradant in the form of wheels (type (c) of Figure 11.4) but both are driven at different speeds to give slip. The Conti is used as another name for the DIN apparatus, although Buist^^ gives it as a separate machine, and the Grasselli is the same as the DuPont. The Martindale abrader is usually seen as a four station machine which uses cloth as the abradant, but coarser and faster acting materials can be substituted. The principal feature of this machine is that the test pieces are rubbed successively in different directions as the motion takes the form of a Lissajous figure. It is mostly used with coated fabrics^^. The Schiefer abrader, which is also known in Britain as the WIRA carpet abrader, is of the form (e) in Figure 11.4. Its principal feature is that it produces a constant relative speed between the test piece and abradant at all points on the test piece, whilst the direction of relative motion changes steadily around a full circle. It is a versatile machine in that a variety of test piece holders can be fitted and the abradant is readily changed, including the use of serrated metal surfaces. It is used for testing polymeric artificial sports surfaces. The Frick-Taber^^ test is a modified Taber, in which loose abradant is introduced between wheel and test piece, used particularly in the evaluation of flooring materials. It is specified in EN 660-2^"^. The Laboratory Abrasion Tester 100 (LAT 100)^^ uses a wheel test piece on an abrasive disk geometry so is the type (d) of Figure 11.4. The abrasive disk is driven and the speed, contact force and the slip angle of the test piece can all be varied. What makes this apparatus so different from all the others is not only the versatility but the sophistication of the instrumentation and the computer control. This means that it can be used to obtain data as a function of several parameters and combine results to make predictions of wear for the extremely complicated service conditions of tyres. Not surprisingly, it is very expensive and unlikely to be used on a routine basis outside of tyre companies. Nevertheless, is has now been proposed for standardization in ISO TC 45. A further attempt to better simulate the wear conditions for tyres resulted in the FKK wear tester^^. An abrasive belt is run over two drums and it is claimed that the same friction conditions as service are achieved. The test

Friction and wear

239

runs under constant braking force and the temperature is controlled by cooling fluid in the drums. Nah et al^^ developed a modified form of blade abrader which is similar in concept to the Pico abrader but covers a larger area to give better sensitivity. The development included an arrangement to give constant frictional torque and the movement of the knife is suggested as an alternative to weight loss for measuring abrasion. A number of tests have been developed with particular products in mind. The use of loose abradant to impinge on the rubber^^' ^^ is particularly suited to surface coatings and a reciprocating test specifically for coatings has also been described^^ Loose particle tests and tests to simulate wear by slurries have application for pipes and conveyors^^'^"^ and there is an ASTM publication on wear by slurries^^ The NBS test^^ is specifically for sole and heel materials and there are a number of other tests used in the footwear industry. A good example of a specialised approach is a machine for 0-rings^^ where the abradant is a roughened metal disc which is immersed, with the 0-ring, in an abrasive fluid such as drilling mud. Abrasion of rubber by loose abradant is clearly a different situation to that where solid abradants are involved. The mechanisms of wear under those conditions have been investigated for several materials^^'^^. Although it is not really a different method, it should be noted that to obtain correlation with service conditions where wear rate is low, very tiny quantities of material lost have to be measured. These small losses may be difficult to measure by the usual weighing or dimensional methods and radioisotope techniques have been used, as for example in the method of Patel and Deviney'^^. Conceivably, a method for measuring very low levels of wear on a tyre by the change in reflectance of an applied paint^^ could be used in laboratory tests. The temperature during test is often not known. A comprehensive investigation of the temperature for the test wheel in a modified Lambourne abrader has been made by Ramakrishnan et al^^ using an infrared camera. The predicted rise for thermoplastics in a pin on disk system has been compared with measurements^^ and, presumably, the calculation could be applied to other circumstances.

2.7

Comparison of Methods

It is to some extent pointless to compare abrasion testers except in the context of their correlation with a particular product and service condition. If a general comparison is attempted, this will inevitably be subjective. However, it can be commented that the DIN abrader has become increasingly popular and is very convenient for routine control use. The

240

Physical testing of rubber

Akron and the improved Lamboume are distinctive for the abihty to vary shp angle in a simple manner and the Schiefer for giving uniform multidirectional abrasion. Probably the most versatile commonly used apparatus is the Taber because of the very wide range of abradants readily available and its ability to operate with lubricants. The LAT 100 is in a class of its own as regards sophistication and features. All of the abraders developed for rubber testing, the Akron, DuPont, Dunlop etc, were primarily aimed at testing tyre compounds. Noboru Tokita at al^^ have discussed tyre wear testing and point out that it is virtually impossible to simulate the total wear pattern and to determine tread life from laboratory abrasion testers, but many people have tried. The LAT 100 approach using multi conditions would seem to stand the best chance. Published accounts of studies of correlation of laboratory abrasion tests with service are not abundant. The study of wear of flooring materials^^ previously mentioned is comprehensive and the same subject has been considered by Gavan^"^. Three abrasion machines were used by Satake et al^^ to study the correlation with tyre wear, and an example of tests using the constant power principle versus tyre wear is given by Powell and Gough^^. Discussion of correlation with tyre wear and of abrasion with other physical properties is contained in Buist's paper^^. Krishnan et al^^ investigated the relationship between laboratory test conditions and road wear results and found that ranking of compounds changed between blunt and sharp grinding wheels. Moakes^^ studied several machines in relation to wear trials on footwear compounds, and work at SATRA and CST^^ used PVC and microcellular materials as well as rubber. Borrof*^ investigated the validity of the NBS test for footwear and Dickerson^^ briefly discusses the conclusions gained from extensive comparative tests on solings and bottoming materials at SATRA. (The results are given in SATRA Internal Reports). Magomedov^^ devised a method to relate abrasion resistance with other properties of soling materials and established a relationship between abrasion resistance measured on a Poznac instrument and sole wear resistance. The reliability of the Taber abrader has been studied by Hill and Nick^^ and its use with polymers discussed by Brown and Crofts^^. The DIN machine has been compared with the Akron and Taber method^"^ for general use with rubbers. Several papers in an ASTM publication'^^ consider the performance of various tests to assess coatings, and the use of a metal mesh abradant to assess the wear of conveyor belts is given by Polunin and Gulenko^^

Friction and wear

241

REFERENCES
1. James D I. RAPRA Members Journal, July 1973, pi70. 2. Ohhara R. Int. Polym. Sci., 23, 6, 1996, pT25. 3. Nakajima Y, Takahashi F. Int. Polym. Sci., 30, 2, 2003, pT36. 4. James D I. J. Sci. Instrum., 38, 294, July 1961. 5. Griffin G J L. Wear, 17, 399, 1971. 6. Mustafa M R and Udrea C. Plast. Mod. Blast., 21,10, 1969, pi 14. 7. Bailey M W and Cameron A. Wear, 21, 1,43, 1971. 8. West G H and Senior J M. Wear, 19, 1.37, 1972. 9. Jost H. Plaste u Kaut, 16, 3, 175, 1969. 10. Wilson A and Mahoney P. Rubb. World, 165,6, 36,1972. 11. Benabdallah S M H. Polym. Test., 9, 3, 1990, pl95. 12. Bielinski D, Janczak K, Janczak T and Slusarski L. Polim. Tworz. Wielk., 36, 10, 1991 p380. 13. Roberts A D, Alliston-Greiner A F. ACS Rubber Division Meeting, Denver, 18-21'* May 1993, paper 61. 14. Bandel P, Di Bernardo C. ACS Rubber Div, Toronto, 1991, Paper 70. 15. James D I and Newell W G. Polym. Test. 1, No. 1, 1980. 16. James D I and Mohsen R. Polym. Test, 3, No. 2, 1982. 17. Giles C G, Sabey B E and Cardew K H F. Road Research Technical Paper No. 66, HMSO, 1964. 18. Malkin F and Harrison R. J. of Physics D, 13, L77-79, 1980. 19. Brown R P. Polym. Test., 7,4, 1987, p279. 20. Wilson M and George M. Satra Bulletin, 1988 21. Wilson M. J. Test. Eval., 24, 6, 1996, p377. 22. James D I. Polym. Test., 18, 1, 1999, p3. 23. James D I. Rubb. Chem. Technol. 153, No. 3, 1980. 24. James D I. Ergonomics, 26, 1, 1983, p83. 25. Marpet M. J. Test. Eval., 24, 4, 1996, p245. 26. Marpet M I, Fleischer D H J. Test. Eval., 25, 1, 1997, pi 15. 27. Slipping-Towards Safer Flooring, Proceedings of a seminar, Shawbury, 29* Sept. 1994, Rapra Technology Ltd., Shawbury, UK. 28. Rowland F J, James D I, Hughes R C. Polymer Testing '95, 15-16* June 1995, Shawbury, Paper 17. 29. Hughes R C, James D I. Polym. Test., 21,1, 2002, p39. 30. Mitsuhashi K. Int. Polym. Sci. Technol, 16,2, 1989, pT31. 31. Sviridyonok A I, Kirpichenko Y E. ACS meeting St. Louis, 1984, Proceedings, p333. 32. Robinson W H and Kopf R E. Mater. Res. Stand., 9, 7, 1969, p22. 33. James D I. Polym. Test., 5, 6, 1985, p403. 34. Kulakowski B T, Buczek F L, Cavanagh P R, Pradhan P J. Test. Eval., 17,4, 1989, p234. 35. Wada N, Inoue A, Kurita Y, Nakajima M, Nojima Y, Okubo K, Kawasaki I, Jonen H, Sugari M, Uchiyama Y. Int. Polym. Sci. The., 24, 3, 1997, pT51. 36. Wada N. Int. Polym. Sci. Tech., 25,2, 1998, pT52. 37. ISO 15113,1999. Determination of frictional properties. 38. BS 903 Part A61, 2000. Determination of frictional properties. 39. BS 3424 Part 10,1987. Determination of surface drag. 40. BS EN ISO 8295, 2004. Plastics film and sheeting - Determination of coefficients of friction. 41. James DL Polym. Test, 8, 1, 1989, p9.

242

Physical testing of rubber

42. James D I. Chem & Ind., No. 3, 1987, p89. 43. Lancaster J K. Plastics and Polymers, Dec. 1973. 44. James D I (Ed.). Abrasion of Rubber, Maclaren and Sons, 1967. 45. Gent A N and Pulford C T R. J. Appl. Polym. Sci. 28, No. 3,1983. 46. Schallamach A. Int. Polym. Sci. & Technol. 6, 1, 1978, pT44. 47. Shallamach A. Int. Polym. Sci. & Technol, 6, 1, 1978, pT63. 48. Shallamach A. Progress in Rubber Technology, Elsevier, 1984. 49. Thavamani P, Bhowmick A K. Rubb. Chem. Technol.,67, 1, 1994, pi29. 50. Ahagon A. Int. Polym. Sci. Technol., 23, 6, 1996, pT32. 51. Lancaster J K. Wear, 14, 223, 1969. 52. Powell E F and Gough S W. Proc. 3rd Rubb. Tech. Conf., Cambridge, 1954, p.460. 53. ISO 23794,2003. Abrasion testing - Guidance. 54. ISO 4646, 2002. Determination of abrasion resistance using a rotating cylindrical drum device. 55. BS 903:Part A9, 1988. Determination of abrasion resistance. 56. ASTM D5963, 1997 (2001). Abrasion resistance (Rotary drum abrader) 57. ASTM D2228,2002. Relative abrasion resistance by the Pico abrader method. 58. ASTM D3389, 1994. Coated fabrics - Abrasion resistance (Rotary platform, double head abrader). 59. ASTM D1630, 1994 (2000). Abrasion resistance (Footwear abrader). 60. Buist J M. Trans. IRI, 26, 192, 1950. 61. Frick O F V. Wear, 14, 2, 1969, pi 19. 62. ISO 5470-2 Rubber or plastics coated fabrics - Determination of abrasion resistance Martindal abrader. 63. Bradway D, Irvin W E J. Test Eval., 17, 2, 1989, p82. 64. EN 660-2 Resilient floor coverings - Determination of wear resistance - Frick-Taber test. 65. Grosch K A. TyreTech conference, 1998, Paper 14. 66. Koike A, Kobayashi I, Furuta I, Yamazaki S. Int. Polym. Sci. Technol, 29, 5, 2002, pT27. 67. Nah C, Cho J, Kaang S. Polym. Test., 17, 1, 1998, pi 1. 68. Tilly G P. Wear, 14, 223, 1969. 69. Egorov B N, Poganov E V and Lyashevich V V. Plast. Massy., 10, 65, 1971. 70. Lancaster J K. ASTM Special Publication 769, 1981. 71. Abu-Isa I, Jaynes C B, School R. Rubb. Plast. News, 16,4, 1986, p30. 72. Arnold J C, Hutchins IM. J. Nat. Rubb. Research, 6,4, 1991, p241. 73. Lowe D, Marshall G P. Plastics Pipes, York, Proceedings, 1985, p8. 74. Kommann M, Genet M, Anderson E. Internal & External Protection of Pipes, Nice, Proceedings, 1985, pi 79. 75. ASTM Special Publication STP 946, 1987. 76. Burr B H and Marshek K M. Wear, 68, No. 1, 1981. 77. Zhang S W, WangD, Yin W. J. Mat. Sci., 30, 18, 1995,p4561. 78. Kopchenkov V G. Int. Polym. Sci Technol, 31, 3, 2004, pT24. 79. Patel A C and Deviney M L., Rubber Division of the American Chemical Society Meeting, Detroit, May, 1973, p787. 80. Chung S R, Lee K S, Yu H J. Tire. Sci. Technol, 26,2,1998, pl20. 81. Ramakrishnan R, Kliman J, Donovan J A. ACS Rubber Division Meeting, Philadelphia, 2-5*^ May 1995, Paper 43. 82. Kar M A. Wear, 67, 1981, p71 83. Tokita N, Shieh C-H, Ouyang G B, Patterson W J. Current Developments In Product Manufacture and Applications, Kuala Lumpur, Proc, 1993, p37. 84. Gavan F M. Mater. Res. Stand., 10, 7,24, 1970.

Friction and wear

243

85. Satake K, Sone T, Hatnada, M and Hayakawa K. J. IRI, 4, 5, 1970, p217. 86. Krishnan V, Ramakrishnan R, Donovan J A. ACS Rubber Division Meeting, Philadelphia, 2-5* May 1995, Paper 44. 87. Moakes R C W. J. IRI, 5,4, 1971, pl55. 88. Maddams J S. SATRA Bulletin, Dec. 1971. 89. Borroff J D. Elastomerics, 109, No. 12, 1977. 90. Dickenson K. SATRA Bulletin, May 1984. 91. Magomedov S S. Int. Polym. Sci. Technol., 30, 10,2003, pT29. 92. Hill H E and Nick D P J. Paint Tech., 31,494, 1966, pi23. 93. Brown R P and Crofts D. RAPRA Bulletin, Aug. 1971. 94. Brown R P. RAPRA Members Report No. 17, 1978. 95. Polunin V T and Gulenko G N. Kauch. I. Rezina, 31,7, 1972, p37.