You are on page 1of 29

A Simple Theory of Static and Dynamic Hardness Author(s): D.

Tabor Reviewed work(s): Source: Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences, Vol. 192, No. 1029 (Feb. 4, 1948), pp. 247-274 Published by: The Royal Society Stable URL: http://www.jstor.org/stable/98140 . Accessed: 03/02/2012 10:40
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

The Royal Society is collaborating with JSTOR to digitize, preserve and extend access to Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences.

http://www.jstor.org

A simple theory of static and dynamic hardness
BY D. TABOR, Research Group on the Physics and Chemistry of Rubbing Solids,

Department of Physical Chemistry, University of Cambridge (Communicated by Sir Geoffrey Taylor, F.R.S.-Received 17 May 1947)

When a hard spherical indenter is pressed into the surface of a softer metal, plastic flow of the metal specimen occurs and an indentation is formed. When the indenter is removed it is found that the permanent indentation is spherical in shape, but that its radius of curvature is greater than that of the indenter. It is generally held that this 'shallowing' effect is due to the release of elastic stresses in the material around the indentation. It is clear that if the recovery is truly elastic it should be reversible and that a second application and removal of the indenter under the original load'should not change the size or shape of the indentation. Experiments show that this is the case. This means that when the original load is reapplied, the deformation of the indenter and the recovered indentation is elastic and should conform with Hertz's equations for the elastic deformation of spherical surfaces. Measurements show that there is, in fact, close agreement between the observed deformation and that calculated from Hertz's equations. These results have been applied to the case of indentations formed in a metal surface by an impacting indenter. The energy involved in the elastic recovery of the impacting surfaces is found to account for the energy of rebound of the indenter. This analysis explains a number of empirical relations observed in dynamic hardness measurements, and, in particular, reproduces the calibration characteristics of the rebound scleroscope. The results also show that for very soft metals the dynamic hardness is very much higher than the static hardness, and it is suggested that in rapid deformation of soft metals, forces of a quasi-viscous nature are involved. In the third part of the paper a simple theory of hardness is given, based on the theoretical work of Hencky and Ishlinsky. It is shown experimentally that for a material incapable of appreciable work-hardening, the mean pressure Pm required to produce plastic yielding is related to the elastic limit Y of the material by a relation Pm = cY, where c is a constant having a value between 2-6 and 3. An empirical method is described which takes into account the work-hardening produced in metals by the indentation process itself. This results in a general relation between hardness measurements and the stress-strain characteristic of the metal, and there is close agreement between the theory and the observed results. In addition, the theory explains the empirical laws of Meyer.

1. STATIC HARDNESS MEASUREMENTS

The hardness of a metal is often defined as its resistance to indentation (O'Neill I93 4) In the Brinell hardness test (Brinell i9oo; Meyer I908), a hard steel ball is pressed under a fixed normal load on to the smooth surface of the metal to be tested. When equilibrium has been reached, the load and indenter are removed, and the diameter of the permanent impression measured. The hardness is then expressed as the ratio of the load to the curved area of the indentation (Brinell hardness) or as the ratio of the load to the projected area of the indentation (Meyer hardness). In both cases, the hardness has the dimensions of pressure. The relation between load and size of indentation may be expressed by a number of empirical relations. The first of these, known as Meyer's law, states that if F is the load applied and d the diameter of the impression left when the indenter is removed (1) F = kdn, [ 247 ]

for hard metals. d3.D2. the values of these constants change. For balls of diameters D1.dn3. if the indenter is replaced in the recovered indentation and the original load applied. This effect has generally been ascribed to the release of elastic stresses in the metal specimen.3 -* (3) where A is a constant... For completely unworked materials. the diameter and curvature of the recoveredindentation should be unchanged. giving impressions of diameters d1. Over a wide range of experimental conditions it is found that F = kd2. using loads ranging from250 to 3000kg. A few of these values were compared with those obtained from direct photomicrographsof the sections across the diameter of the .(b)a metallographic section across the diameter of the indentation. but that its radius of curvaturemay. The radiusof curvaturer2of the recovered indentationwas also measured.. Tabor where k and n are constants for the material when the diameter of the ball is fixed. The power of d is fixed. That is to say.D3.. a series of relations is obtained of the type F = kldln = k2dn2= k3. (2) In a very extensive series of investigations Meyer (1908) found experimentally that the index n was almost independent of D but that k decreased with increasing D ( in such a way that A = kDn-2-k Dn-2 _k Dn-2 -11 2 2 -3 . When balls of different diameters are used. . the surfaces should deform elastically.248 D. be as much as three times as large as that of the indenting sphere. When conical or pyramidal indenters are used as in the Ludwik and Vickers hardnesstests respectively a simplerrelationshipis observed. A series of impressionswere made with hard steel balls of various diameters on various metal surfaces. It is at once evident that if the recovery observed is an elastic one. Experiments were carriedout to test this.0 (4) for an indenter of fixed angle. whilst for fully worked materials it is close to 2. It has long been known that the permanent indentation left in a metal surface deformed by a hard steel ball has a larger radius of curvature than that of the indenting sphere. . The values of d were reproducible to less than 1 %. but little work of an analytical nature has been carriedout to relate this to the elastic propertiesof the metal and the ball. it should be essentially reversible. and on removing the load. The diametersd of the impressionsformedwere measuredafter 1. The radii of curvature as determinedby the profilometermethod were reproducibleto about 4 %. 2.d2. 3 and 5 cyclic applicationsof the load. n has a value near 2 5.using (a) a delicate profilometer. Some very careful measurements by Foss & Brumfield (I922) have shown that the indentation is symmetrical and of sphericalform. but k depends upon the angle of the cone or pyramid used. The value of n is generally greater than 2 and usually lies between 2 and 2-5..

585 0*200 0. and a normal force of F dynes is applied.20 -.595 0.55 0-203 0. it is possible to apply the classical laws of elasticity to this portion of the deformation process. the agreement was of the order of 1 to 2 %.330 1'37 0*366 5*54 3 0.270 1.39 0 370 5 60 2 0.952 0. ball) the radius of curvature of the indentation by the profilometer method (mean of three determinations) was 0*595 cm.652 0*338 1F37 0 365 5. (cm.263 1.) 500 500 500 1000 3000 3000 dimensions of indentations (cm. both surfaces are elastically deformed to a common radius of curvature r where r2 > r > r1 and the deformed surfaces finally .19 0. When a hard steel sphere (radius of curvature r1) is placed in the indentation.16 0d186 0. This shows that the deformation occurring at the final stage of the original deformation is reversible and is therefore essentially elastic.677 03330 1. These values are typical. The results are given in table 1.41 d l k - < I * I d .202 0.677 0. on mild steel (10mm.68 0. and it is assumed that it consists of a plane surface XA BY containing a depression of spherical form of radius of curvature r2 and of diameter d = 2a (figure 1 a). and by the direct contour method 0 605 cm.270 1*21 0.270 1.15 0-183 0.16 04185 0.260 1. The condition of the surface of the metal after the indentation has been formed is idealized.27 1*21 026 1.lI d-j (a) FIGURE 1 (b) It is seen that the indentation remains essentially unaltered in diameter and curvature after the second and third applications of the original load. For example. Since the 'recovery' of the indentation is truly elastic.) number of applications of load - metal brass aluminium alloy mild steel hardened steel d r2 d r2 d r2 d r2 d r2 d r2 1 0. with a single application of load of 500 kg.952 05 05 0*952 1'59 l load (kg.E 0-192 _ 0*206 0.A simple theory of hardness 249 indentation.338 1F31 0 366 5 68 5 0. TABLE 1 radlus o ball r.*1 .) 0.

the relationship between d. The results are given in table 2. the same equation will result. It is assumed that E2 = 20 x I10 dynes/cm. the surface rises at the regions A and B as in figure 2 a or falls as in figure 2 b.) 017 500 aluminium alloy mild steel 0*952 0Q5 0. It is assumed that there is very little change in the diameter d during this deformation. E2 are Young's moduli for the indenter and the metal. Tabor touch over the boundaries of the indentation (figure 1 b). It is seen that the agreement between the last two columns is reasonably good. x d " A d (a) FIGURE 2 (b) This equation is applied to the previous measurements of rl. In a discussion of the derivation of this equation. use is made of the observations of earlier workers. according to Hertz's classical equations (Hertz I896) describing the elastic deformation of spherical surfaces.) load (kg.250 D.39 0 202 0*330 0*22 0*36 3000 1P59 2. If. Then. for example.2 for .5 0. provided the projections or depressions at A and B are not too marked. r2. and where a value of 0 3 for Poisson's ratio has been assumed. F and d. The value of d obtained from equation (5) is compared with the observed value of the diameter of the recovered impression.952 0-5 1P21 0*66 1-15 0 605 0*27 0-178 0*26 04183 0-26 0-18 0-30 0. an assumption which is generally accepted as being valid to within a few per cent. r1 and r2 is given by d = 2a = 2-22[ I2 r-rEl E2)] +5) where E1. Prescott (I927) has indicated that even if the surface XABY is not a plane. (1) Profiles given by Batson (I9I8) for a ball of diameter 1Omm.) 250 material brass ri 0 5 r2 0*64 d 0-160 calculated d (cm.80 0-370 0*39 As a matter of interest. on three types of steel.2) 10 x 1011 observed values (cm. particularly as the accuracy in determining r2 is not better than about 4 ? TABLE 2 assumed value of E1 (dynes/cm.952 0 677 1. the above equation is still valid.20 7 x 1011 20 x 1011 250 500 500 hardened steel 20 x 1011 1000 3000 0. and a load of 3000 kg.

The results are given in table 3. TABLE 4 assumed value of E2 load observed values I`_calculated (dynes/sq.) 043 0*37 0.5 0-5 r2 (cm.555 0 330 0 497 0.276 0-499 0. whilst the agreement for the higher loads (3000 kg.A 8imple theory of hardness 251 all the steels. diameter and 20 x 1011dynes/cm.5 0.) 0*518 0. diameter on various brasses.324 0 407 0*445 calculated d (cm.302 d (cm. It is seen that the agreement between the observed and calculated values of d is reasonably good.r1) may be very large.) is good. This is probably because for soft metals the values of r2 are very little different from r1 for high loads. so that the errors introduced in calculating r1r(r2 . (cm.28 (3) Profiles given by Foss & Brumfield (I922) for a ball of 10mm.) soft brass 1 soft brass 2 hard bronze 3 soft bronze 4 hard bronze 5 soft bronze 6 9 x 9 x 7 5x 7 5x 75x 75 x 10"l 101" 101" 1011 101" 1011 (kg.) 3000 500 3000 500 3000 500 r. a load of 3000 kg. The results are tabulated in table 4. on various steels.) 0'5 0*5 0. It is seen that the agreement for the smaller loads (500 kg.) 0.566 d (cm.5 0.) is poor.301 0*240 0 349 calculated d (cm.) 0.32 0-28 0-33 0*23 0'36 .) 0.5 0.56 0593 0*701 1*03 0652 0*814 0*644 1-372 0-568 d (cm.2 for all the steels.521 0.cm.527 0 557 0-531 0.5 0*5 0-5 0-5 r2 (cm.25 0. TABLE 3 observed values material Ni-Cr steel manganese steel rail steel r1 (cm.) 0.) 0*34 0-41 0 49 (2) Profiles given by Foss & Brumfield (I922) for a ball of 1Omm.627 0-569 0-537 d (cm.38 0-66 0-28 0-63 0. It is assumed that E2= TABLE 5 observed values metal 0*5C-A 0*5C-O 0*5C-T 0*5C-W 0*5C-T 0*9C-T 0*9C-O 0*9C-W MKD 455 r. (cm.) 0*440 0*342 0*345 0*26 030 0-31 0. In table 5 results are taken where r is greater than 0(55 cm.) 0*7 0. and it is seen that the agreement between d (calculated) and d (observed) is close.) 0-5 05 05 0-5 0'5 05 r2 (cm.5 0. The results are given in table 5.31 0.) 0.

the last stage of the indentation process is reversible and may be expressed in terms of the elastic constants of the materials and the sphere. plastic flow has ended and the whole of the load is borne by elastic stresses in the material. be put forward. There is a further increase in work hardening and the process continues until the stresses are distributed over a larger impression. a release of elastic stresses. and the results show that on the whole there is good agreement between the observed and calculated values for a wide range of materials and experimental conditions. it rebounds to a certain height and leaves an indentation in the surface. This result may be interpreted as implying that the metal offers a constant pressure of resistance to the indenter equal numerically to the ratio (energy of indenter)/(volume of indentation). The stresses. Vincent (I900) and other workers have confirmed this experimental observation. true that as the calculation of d involves a cube root. As the metal is displaced. the load is removed. and if the same load is reapplied. as we have seen. the agreement is consistent for a great diversity of materials and experimental conditions. therefore. At the end of the process. Nevertheless. If it is further increased. for example. particularly when r2 is not too close to r1. The elastic stresses now reach the limits that the deformed material around the impression can stand. and so fall again within the increased elastic limit. Tabor These results show that in general the agreement between the observed and calculated values of d is reasonably good. the stresses exceed the elastic limit and further flow of the material occurs. there is. This is shown by the fact that after several cyclic applications of the original load the recovered indentation remains essentially unaltered in diameter and in radius of curvature. soon exceed the elastic limit of the metal and plastic flow occurs. If. r1.252 - D. i. however. when the elastic 'recovery' is marked. DYNAMIC HARDNESS MEASUREMENTS If a hard steel or diamond indenter is dropped on to a metal surface. When the ball presses on to the surface. the description in Carpenter & Robertson (I936)). for example.e. of course. therefore. the values of F. work-hardening occurs and the elastic limit of the material increases (see. It is. the metal is first deformed elastically. the volume of the indentation so formed is directly proportional to the energy of the indenter. 2. . Martel (I895) showed that over a wide range of experimental conditions. there is elastic 'recovery'. r2. the surfaces deform elastically until they just fit over the diameter of the original impression. This process continues until the stresses are now distributed over an impression of such dimensions that the stresses are within the elastic limit of the deformed material. If the load is removed or reduced. E1 and 12 are not very critical. The following mechanism for the processes involved in the Brinell hardness test may. Discussion The experiments show that when an indentation is formed in a metal surface by a hard spherical indenter. Further. the extent of the elastic 'recovery' of the indentation may be calculated on the basis of Hertz's equations.

If the volume of the remaining permanent indentation is V. This process continues until the indenter is brought to rest.and is referredto as the dynamic hardness number. the energy and All of rebound W2. (i) When the indenter first strikes the metal surface. the height of rebound is roughly proportional to the static hardness of the material concerned (Shore I9I8).A simple theory of hardness 253 This ratio has the dimensionsof pressure. (ii) Plastic deformation of the metal now occurs accompanied by a building up of further elastic stresses in both the indenter and the metal. In a detailed investigation of the impact of spheres of similar metals. there is a residual elastic deformation which is reversible. that remains thereforeis to calculate W2 the volume Vr. In the following analysis. a differentprocedurewhich considerably simplies the problem will be adopted. Andrews (1930). however. calculated the time involved in each of these portions of the impact.. who was mainly concerned with the period of the collision. plastic flow occurs. This assumption implies that whenever the pressure during impact reachesthe value P. Later workers have discussed the validity of this relation in some detail. The analysis applied to the last two parts of the process was. It is assumed that the mechanism involved in the dynamic indentation is essentially the same as that which occurs under static conditions. It is assumed that the energy . In what follows the processes involved in impact experiments will be analyzed. Now consider the indentation after impact has occurred. the work done as plastic energy in producing this indentation is by definition of P given by (6) Clearlythe energy 11 is the differencebetween the energy of impact W. It is assumed that there is a dynamic yield pressureP which to a first approximation is constant and which is not necessarilythe same as the static pressurenecessary to cause plastic flow. of an admittedly approximate nature. when the plastic deformation has been completed. where one is concernedessentially with the forces involved and not the time of collision. After the collision the indenter rebounds to a height 2 and leaves a permanent indentation in the metal surface of diameter d =-2a. and in particular it has been suggested that the energy of rebound should be taken into account in calculating the dynamic hardness. and Suppose the indenter has a mass m and a spherical tip of radius of curvature rl and that it falls from a height hAon to a flat metal surface. where the height of rebound is used as a measure of hardness. (iii) There is now a release of elastic stresses in the indenter and the material surroundingthe indentation. a simple theory which explains a numberof empiricalrelations observedin dynamic hardness measurementswill be developed.elastic deformationtakes place until the mean pressure developed is sufficient to cause plastic flow of the metal. and so long as plastic flow continues the pressure remains constant at this value. The processof impact may be divided into three main parts. That is to say. It is found that if the height of fall is constant. A different approach is that adopted in Shore's scleroscope. as a result of which rebound occurs.

as a result of the elastic deformationof both contacting surfaces.254 D. At any intermediate instant when the region of contact has a diameter 2cc(where ac a) the forcef on the indenter. given by equation (5a). Since there has been a release of elastic stresses in the indentation.27F2 1 energy of rebound. it would deform the indentation (and itself) elastically. is 1 W2= mgh2 = Fa LE1 E21 (8) where a value of 03 is again assumed for Poisson's ratio. r2 Hence 1. and just touch over the diameter d. If a suitable load F were applied to the indenter for a very short interval. Tabor involved in the release of these elastic stresses is equal to the energy of rebound of the indenter. Then the integral of fdz over the range ca= 0 to ac= a is the total elastic energy stored in the surfaces. the centre of the indenter has descended a distance z (Prescott 1927) given by Z3f [ 1-o-21--2] (7) where o-. The volume Vrof the permanent indentation left in the surface may be written to a first approximationas r= lTa4/4r2. consider the indentation after the impact has occurred. where again d = 2a = 2. which we equate to the o. its radius of curvature will not be r1 but will be somewhat greater. is < f F a 3. = (5b) At this stage. Finally.0-O68- r1 i-+--a3LE1E+2 (5 c) W F2 F 1 11 7Ta4 -P 0-017-I 4r1 a L1E2j(9 . Hence P W3 Wl-W2 = PVr= 4r2 (6a) We now express r2in terms of r1 and F from equation (5a).and o2 are Poisson's ratio for the indenter and the anvil respectively. it is assumed that Young's moduli for the indenter and the metal are essentially the same as for static conditions. Again. As the indenter sinks into the indentation the force increases from zero and reaches the final value F given by equation (5a) as the full contact across the diameter d = 2a is completed. A simple integration shows that this energy.22[ir1 2r2 /r1E + E (5a) The elastic energy involved in this processis estimated by calculatingthe external work performedin pressing the indenter into the indentation. say r2.

In particular.mgh1/V0.Then the work W3expended in displacing plastically a volume Vrbecomes _3?PE~ (6a) W3 n+ 2 r. P will tend to increase during impact in a manner similar to that observed in static hardness measurements. to express this effect quantitatively.(a where P is now the mean pressure at the end of the deforming process. The second term. as described in equation (1).A simple theory of hardness 255 since the force F at end of indentation is equal to PnTa2. however. Hence ( 0) Thus p= m( V8-2) (11) The validity of this analysis depends on the assumption that the internal forces occurring in the actual impact are essentially the same as those involved in the analytical model just described. P . This will tend to increase P at the initial stages of the deformation when the velocity of displacement is a maximum (see later). i92. Effect of variation in the value of P In the above derivation it has been assumed that P is a constant throughout the process of impact. Some estimate of the order of this effect may be made by assuming that. where VI is the apparent volume of the indentation which would be obtained if the indentation was considered to have the same radius of curvature as the indenter. The second reason is that work hardening of the deformed material will occur during the formation of the indentation. on analogy with the static indentations. It is also assumed that the temperature rise of the material around the indentation during the impact is small and has a negligible effect on the strength properties of the metal.h2)/Vj. we obtain rn~ p-n+2 2n . . This is also the pressure involved in the calculation of the rebound. Substituting this value of W3in the appropriate equations. two reasons why P may be expected to vary during the collision.1I 2(n?+ 2)21) a 4 Vol. it is assumed that the elastic waves set up in the indenter and the metal specimen absorb a negligible amount of energy. we can write P=kdn-2 (l a) where n lies between 2 and 2X5. (lla) '7 A. however. It is difficult. It is at once apparent that if the rebound is not very large (so that h2 is small) the results will not be very different from the equation given by Martel. The first term of equation (9) is simply PV. nor from the equation suggested by later workers. The first is a dynamic effect associated with the kinetic displacement of the metal during impact. is seen to be equalto 5W2. P = mq(hl . As a result. There are. by comparison with equation (8).

The validity of equations (8) and (11) Remembering that at the end of the indentation process. i. straight lines should be obtained with a slope of . so that this term tends to give lower values for P. will never be more than about 10 O. On the other hand. If P is not constant but varies in the manner given by equation (la) it is found that the logarithmic graph of h2 against Va is still a straight line. the term in front of the main bracket increases from 1 to 1P12 as n increases from 2 to 2 5. this means that h2 is proportional to Vla for any fixed material. and that the slope lies between 0 7 and 0-85. Plotting A2 against V"on logarithmic ordinates./ mild ~~~~~~~~~~~~ ~~~o steel (3~~~~~~~~~~E ni-cr steel 0 aluminium 7 0 x_ ctheoretical slope Q002 00 5 -0010x05 0 02 0 10 2ai apparent volume of indentation FIGuRE 3 (cubic units) . The difference. but the slope has a value of 1(3 + 2n . and it is seen that this is approximately true. 20 X 0 / 0 nickel. equation (8) as 27a3P2 1 11 we rewrite h= mg L E2j 8a Since the apparent volume of the indentation Vais proportional to a4.4). however. It is seen from figure 3 that in fact the points for each material lie on a straight line.e.h2as n varies from 2 to 2-5. Tabor The last term in the bracket varies from 3h2to 4. The total effect is to give values of P which are somewhat greater than those given in equation (11). if P is constant. F -ra2P. Some results taken from Edwards & Austin's paper (I923) are plotted in figure 3. | 2- / / - / x / / . it varies from to 1 as n varies from 2 to 2 5.256 D.

a straight line of slope I.and P is given by P5= h- )3- m3[ /E I/E2]4 (12) (A similar relation is obtained if equatioD (II1a)is used instead of equation (II). These curves reproduce in fact the main characteristics observed in the practica. is almobst directly proportionalto the dynamic yield pressure.1 If allowancei? made for the 'factthat softer metals usually have a smallerYoung's modulus. the values of P being calculated according to equation (11). It is also evident that -over a wide range of experimental conditions the height of rebound. any dependence of P on the size of the indentation does not appreciably affect the relation. for a given heigaht fall Chl. this factor may be treated as a constant and P may be plotted as a function of h2 of The theoretical curve is shown in figure 5.should be obtained. therefore.. the slope of the straight line being 0-51.A simple theory of hardness Again.tshown in the dotted curve.) Since the bracket involving Young's moduli does not vary greatly for most metals.. for a fixed height of fall. If. a may be eliminated between equations (8a) and (11). In this case. Results taken from the same paper are plotted in figure 4. h2 is plotted against 257 of a fixed diameter. h.l calibration of rebound sclerescopes. 4 h2 (cm-) tungsten 20 50 tungsten 100 Finally. the curve is modifiedin a mannersimilarto tha. I7-2 . h2 should be proportional to P1(I/E + I/E2) on logarithmic ordinates. The resulting relation between hl. for indentations P2[1/E1 + 1/E2]. There is again good agreement. 20 0 - 0~~~~~~~ aluminium 200 200 1 coprnickel 01( mild height of rebound FIGURE.

As a matter of interest. If we go back to the original equations and calculate the final average pressure PI developed between the indenter and the anvil in a purely elastic collision.5 1 dynamic yield pressure (arbitrary units) FICuuI2 5 3 follow from this result.. diameter 1 cm. the processes of impact and rebound become entirely elastic.2. I 26 6r1(1/E?+ I/E2)V' (12a) again assunminga value of 0^3 for Poisson's ratio. 20 0-5 2 15 2./mm. is 100 cm.258 D. plastic deformation will not occur if the yield pressure of the specimen is higher than the value of P given in equation (12a). we obtain a relation P I. the pressure obtained from equation (12) is then the final mean pressure between indenter and specimen. In this case. we find that Pe = 400 kg. it is possible to calculate the value of Pe from typical data. equation (12) is valid right up to a rebound of 100 %. First. and there is no plastic deformation of the anvil (see Taylor 1946). Tabor Te condition fo elastic collisions It is interesting to consider what happens when the rebound is equal to the height of fall. and if E1 E2 2 x 1012dynes/cm. If the height of fall h. Two conclusions 4 80 / 60 - 40 a4) . Secondly.2. In the latter case. This means that if the anvil has a yield pressure* less than about * This is the yield pressure for infinitesimal deformation. It corresponds to the transition from elastic deformation to the onset of plastic deformation and is considerably smaller than the yield pressures observed in the fully plastic deformations occuLrring practical hardness in . This is exactly the same as the value for P obtained from equation (12) when h is put equal to hl.. and the indenter is a steel ball of mass 4g.

In a typical case Taylor (1946) finds that for a steel anvil the collisions just remain elastic when hL = h2= 1 cm.2./mm. work hardening will make this difference more marked. A decrease of r by a factor of only 3-2 will double the value of Pe. Although in this case both spheres are plastically deformed at the region of contact. For example. This corresponds to a value of Pe of about 160 kg. In fact. iii. These observations may prove of value in the design of impact hardness equipment./mm. the general form of these curves is fully confirmed in practical experiments.2. there will be no plastic deformation of the specimen.) As will be seen later. A more sensitive method. In this case the rebound method will give the same height of rebound for all metals with a yield pressure higher than 400 kg. had a Brinell hardness number of 351 which means that under practical hardness measurements its yield pressure was over 360 kg. an increase of either of these by a factor of 32 will double the value of Pe. Nevertheless./mm. 0 6. Similar curves have been obtained by Raman (I9I8).2 plastic flow will occur. Coefficient of restitution If the indenter falls with a velocity v. In addition. The way in which e varies with the velocity of impact is shown in figure 6. v2 = 2qh2. 0 8. where curves i. however. The steel./sec. 264). This may be readily achieved by increasing the mass of the indenter or the height of fall. so that some deviation may be expected from these curves in practice. and the height of rebound will be equal to the height of fall. iv and v respectively have been drawn for values of e = 1. the yield pressure of the specimen is above 400kg. so that the ratio e = V2/V1 will not be a constant. 0*4 and 0 2 at an impact velocity of 450 cm. (13) It is clear from this relation that v2 depends on vl. is to decrease the radius of the tip of the indenter. (This corresponds approximately to a height of fall of 100 cm. the height of rebound will be very insensitive to yield pressure above about 350 kg. P is not a constant. however. ii. Typical results obtained for cast steel and drawn brass are shown in the dotted lines. equation (12) gives V2 = k(v2-32)-v. Okubo (I922) and Andrews (I930) in experiments on the impact of spheres of similar metals.2. the relation between v1 and v2 is of the same type as in equation (13).2. and the height of rebound will be less than the height of fall. If./mm./mm. the coefficient of restitution e is defined as e = V2/v11 We may find e from equation (12) by putting v2 = 2qhl. however./mm. on to the surface of the anvil and rebounds with a velocity v2. In this case the yield pressure of the material is given by equation (12). . If it is wished to extend the sensitivity of the rebound method for the measurement of higher yield pressures.A simple theory of hardness 259 400 kg. Assuming that the yield pressure P remains essentially constant. measurements (see p.2. the experimental conditions must be modified to give a higher value for Pe in equation (12a).

They were dropped from various heights h1 on to massive anvils of various metals. At very low velocities of impact. Some static experiments were also carried out on the same surfaces. and there will be a corresponding decrease in the coefficient of restitution. as Andrews (I93I) showed for lead and tin alloys.i -s" 0240 \ iv~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~504 I I I 200 400 FiGURE6 600 800 velocity of impact (cm. As the velocity of impact increases. the pressures developed will be insufficient to cause plastic flow. It is apparent from these equations and from the experimental curves that in general the coefficient of restitution of impacting solids capable of undergoing plastic deformation will not be a constant. This equation gives curves which are similar to those given by equation (13). 10 a0- 0.8 o .260 D. The load L required to give an impression the same diameter as . we obtain v -k(v2 -pvMY where . using the same steel balls.) A comparison of dynamic and static hardness numbers Some impact experiments were carried out with steel balls of diameter 05 and I 0 cm./sec. The collision process will be entirely elastic and the coefficient of restitution will be unity. using equation (11). This occurs even with the softest metals if the velocity of impact is small enough. Tabor If instead of equation (11) equation (11 a) is used to derive the relation between (13a) v1 and v2.8 -(2n . but they are appreciably flatter. the amount of plastic deformation will steadily increase. From these observations the dynamic hardness P was calculated. and the height of rebound h2 and the diameter d of the permanent indentations left in the metal surface were measured.1)/(2n + 4).

since it corresponds to the Meyer static hardness number. TABLE 6 diam.) 0'5 1'0 05 1'0 h./mm. when plastic flow is occurring.057 0. This suggests that in calculating P from the volume of the indentation. part of the energy is used in the viscous displacement of the metal as the indenter sinks into the surface.5 0 1'0 d 0'111 0-133 0. The results are given in table 6. higher the greater the height of fall hl. It is seen that the mean pressure required to produce plastic flow dynamically is higher than that required to produce the same amount of flow statically. Typical results are given in table 7. of ball p = - h2) Pm va surface hard steel aluminium alloy (cm. . We call this Pm. In this way. it will be seen that the yield pressure Pr calculated from the height of rebound is less than the yield pressure P calculated from the volume of indentation. and somewhat greater than the static yield pressure Pm. there is a bulk displacement of metal around the indenter. 324 324 50'4 302 50'4 302 50'5 302 50'5 302 51 313 51 313 50'6 312 50'6 312 h2 176 176 18'5 89 20 94 11 56 12 52 0 7'4 0 6'5 0 1./mm. in general.085 0'113 0'172 0-063 0'092 0'126 0'190 0'110 0'163 0'217 0.0 indium 05 1'0 -. During the formation of the indentation. It is also seen that the dynamic hardness is not constant but is.2) (kg. and the effect becomes more marked if equation (1 Ia) is used to calculate P instead of equation (11). This is what would be expected. If the last three columns are compared. lead and indium. This view is confirmed by a calculation of the yield pressure from the height of rebound.329 0'133 0'200 0'266 0'398 (kg.A simple theory of hardness 261 that observed in corresponding impact experiments was found.2) 280 356 342 290 103 129 105 122 74 98 73 86 8-7 11'0 9'2 10'6 4 5 4 5 80 95 81 100 63 74 59 66 5'5 6'0 58 6'3 1 brass 0'5 1'0 lead 0'5 1. This is particularly the case with the softer metals. We rewrite equation (8a) p2m r h2F1] (8b) 2'7a L1/E1-I-J/E~2_ where the suffix r is added to P to show that it is calculated from the rebound height. a static yield of pressure of the material given by L4/7Td2 was calculated.

All the deformation around the indenter is now of an elastic nature. surface hard steel Mass 0 507 g. lead and indium. so that the kinetic yield pressure P is appreciably higher than the static value Pm..4 1. The dynamic yield pressure P is now very much higher than the static pressure P. There is no further bulk displacement of metal around the indenter. forces of a quasi-viscous nature are involved. the average pressure resisting indentation during the impact itself (P) is always higher than that involved at the last stages of the impact (Pr).5 0085 04163 0200 105 6. 302 313 312 89 7.2 aluminium alloy. Discussion These results show that when an indentation is formed in a metal surface by an impinging indenter. the pressure resisting indentation is greater than that occurring in the formation of a similar indentation under static conditions. kg. at the end of the impact where the elastic compression and recovery take place.262 D.1 1-6 129 8.5 1 E = 1'6 x 1011dynes/cm. This dependence of the yield pressure on the speed of deformation suggests that in the dynamic deformation of metals. where the pressures required to produce plastic deformation dynamically are very much greater .2 conditions should be reversible. Tabor This involves the expenditure of kinetic and viscous energy. E2 = 1 x 1011 dynes/cm. Further. cm.7 5 95 5.2 indium.057 65 82 87 74 98 103 63 74 80 lead. CALCULATED FROM HEIGHT OF REBOUND Diameter of ball 05 cm. This is borne out by the results for soft metals. These results are consistent with the view that the forces required to deform metals plastically are greater. With the soft metals the difference between P and P. and the higher the velocity of impact the greater the pressure of resistance. and no energy is expended in pushing the metal bodily away from the indentation. the faster we attempt to deform them. On the other hand. For the harder metals the values of Pr are less than 10 00 higher than Pm. where the plastic deformation has completely come to an end.2 h. the plastic flow of the material has come to an end. whereas P. E2= 10 x 101"dynes/cm. 324 h2 176 d 04111 Pr 304 P 356 Pm 280 brass. As a result the pressure at this stage may be expected to be essentially the same as Pm' This is seen to be approximately the case.. and any kinetic energy imparted to the material under these TABLE 7.2 50 5 302 50'4 11 56 18'5 0-063 0'092 0. whilst the values of P are 20 to 30 00 higher./mm. DYNAMIC YIELD PRESSURE P. E2= 7 x 101"dynes/cm. becomes very marked indeed. remains relatively close to the static values.

the effective yield pressure PJis very much smaller than the mean dynamic yield pressure P which is involved during the course of the impact itself. the pressures involved (Pr) are only a few per cent higher than those involved in the formation of indentations of the same size under static conditions. where Y is the elastic limit of the material as found by pure tension (of frictionless compression) experiments. and finds that when plastic yielding just occurs the mean pressure on the punch Pm is about 2 8 times the yield stress Y of the material. where c is a constant having a value of about 3. -cY. Hencky has derived an approximate solution for the case of a frictionless cylindrical punch penetrating an ideally plastic solid. that to a first approximation. Some typical results are given in table 8. A SIMPLE THEORY OF HARDNESS In this part of the paper. Hill. appreciable amounts of energy are dissipated as a result of the 'viscous' flow of the deformed material surrounding the indentation. 3. the elastic limit Y being found from careful compression experiments. Finally. The solution of Mises's equation for cases involving twodimensional plasticity has been solved by Hencky (I923). the criterion for plastic flow as proposed by Mises is that (Pl P2 )2+ (P2 -p3)2 + (p3 -pl. P3 are the principal stresses in a solid body.A simple theory of hardness 263 than the static values. where the plastic flow of the metal has come to an end and a regime of purely elastic stresses has been reached. at the end of the indentation process. If P1. The solution when axial symmetry is involved presents serious theoretical difficulties which have not yet been satisfactorily resolved. It would seem that in the deformation of soft metals. A more rigid analysis recently carried out by Ishlinsky (I944) yields essentially the same result. Ishlinsky has also applied his analysis to the case of a frictionless spherical punch penetrating an ideally plastic solid. however. where relatively large volumes of metal are displaced. This model corresponds to the Brinell indentation of a metal which does not work harden. plastic yielding occurs when P. increased slightly with the size of the impression made. an attempt will be made to correlate the hardness of a metal with its elastic limit and with the way in which the elastic limit varies with the amount of deformation to which the metal has been subjected. P2. (14) . The values of P. since at the end of the impact where the work-hardening would be a maximum.)2 _ 2 Y2. Lee & Tupper (I 947) and other workers. and finds that the mean pressure on the punch depends somewhat on the size of the indentation formed. Experiments may be carried out to test this result by using metal specimens which have been very highly worked so that they are incapable of appreciable further work-hardening. This cannot be due to the work-hardening which may occur rapidly during the formation of the indentation. but that to a first approximation it is equal to 2-6Y. It is clear.

whilst the pressure resisting deformation rises rapidly. in the range where the whole of the material around the indentation flows plastically.2) Pm ratio Pm/Y metal tellurium lead copper mild steel 2. Tabor Now consider in somewhat greater detail the range over which equation (14) is valid. However./mm. expect that there exists an average or ' representative' value of the elastic limit which is related to the mean pressure Pm by a relation of the same type as equation (14).8 2-8 The onset of indentation which occurs when PI. the region of plasticity is extremely small and the permanent deformation resulting is also very small.1 88 190 2-9 2./mm. We may. Pm % Y.1 31 65 (kg. the elastic limit Y is a constant and equation (14) is valid. = I . TABLE 8 Y (kg. 344). and this in turn depends on the amount of deformation to which it has been subjected. p. the numerical factor increases 3 slightly with the size of the indentation. the elastic limit depends on the amount of work-hardening which the metal has undergone. At this stage. Consequently. With most metals. the deformation of the surfaces is at first elastic and the stresses ar'e given by Hertz's analysis. as given in equation (14).264 D. there is a very gradual increase in the size of the indentation produced. This transition in the value of Pmfrom 1 1 Y to about 3 Y is part of the intrinsic mechanism of plastic deformation and is distinct from the effects produced by work-hardening. probably on account of the increased confinement of the displaced material (see Bishop. The Mises criterion indicates that the material first exceeds the elastic limit at a region below the surface of contact when the mean pressure Pm reaches a value of about I IY (Timoshenko I934. At this stage. As we have seen. When an indentation is formed by a spherical indenter the material around the indentation will be displaced and in general the elastic limit will be raised. however. Making this assumption a general relation between Pm and the size of the indentation may be derived. A stage is soon reached at which the plastic region extends over the whole of the domain around the indentation. as will be seen below. As the load is increased the region of plasticity grows. Effect of work-hardening If the metal is incapable of work-hardening. however. Hardness measurements are usually carried out at loads well above this point. When a hard spherical indenter is pressed on to the plane surface of an ideally plastic body. the elastic limit will not be constant at every point around the indentation since the amount of deformation or strain will in general vary from point to point. Hill & Mott 1945). . equation (14) may be applied to most practical hardness measurements.2) 6.Y is observed only when extremely refined measurements are carried out under very small loads.

be said that the strain produced at the 'representative' region is simply a function of d/r2. For a fixed metal 80 is constant. (15) where c is a constant having a value of about 3. the amount of deformation of strain at the 'representative 'region will be the same if the grain size of the material is sufficiently small as to be irrelevant. the metal has previously been cold worked. therefore. If the metal is initially fully annealed. A convenient method of measuring the elastic limit of a material is to determine its hardness using a .e. We assume that the elastic limit Y is a single-valued function of the deformation or strain so that Y 0(= ) Thus the 'representative' value of the elastic limit will be given by Y = 0(80 +f(dlD)) Equilibrium is therefore set up when Pm-= co (0+f (d/D)). we may consider it as unworked material that has undergone an initial deformation or strain 80.A simple theory of hardness 265 Suppose the indentation has a diameter d and a radius of curvature r2. a smooth single-functioned curve should be obtained for all the loads and ball diameters used. Then for all indentations for which d/r2 is the same. Since r2 is usually very near the radius of the indenting sphere (r1 = D/2) the deformation becomes approximately equal to 81= f(d/ID). If. Further. It is seen that all the results lie about a smooth curve for ball diameters ranging from 1 to 30 mm. the curve is of the samp type as the elasticlimit/deformation (i. This equation means simply that geometrically similar indentations produce similar strain distributions. we may. add this strain to that produced by the indentation. As we shall see later. Hardness as a function of the stress-strain characteristic A more quantitative connexion between the mean pressure and the elastic limit of the material around the indentation may now be considered. It may. Some results by Krupkowski (I 93 I) for annealed copper are plotted in figure 7. however. to a first approximation. Since it is a portion of a sphere its shape is completely defined by the dimensionless ratio d/r2. this is the total strain produced by the indenting process. the stress-strain curve) for annealed copper. Hence the total strain 8 produced at the 'representative' region of the indentation will be given by d = 6' +f (d/D). Co-ordination of results The first result that follows from equation (15) is that we may at once co-ordinate hardness measurements made with various loads and ball diameters on a given metal specimen. so that if Pm is plotted against d/D.

the deformation is expressed as the change in length divided by the compressed length and corresponds to the fractional increase in the area of cross-section of the specimen. At the edge itself the elastic limit rises rapidly and then falls somewhat as the centre of the indentation is . and the elastic limit at each stage of compression determined. This relation may then be used to determine the elastic limit of any specimen of the metal.0 . the amount of deformation or strain and the elastic limit at any stage. Typical results for indentations of various sizes in mild steel are shown in figure 9. one hardness of a muetal obtains a direct relation between the hardness number. one measures the Vickers that has been compressed or elongated by various amounts. since the indentation is geometrically similar whatever its size. Blocks of these metals were carefully compressed by various amounts. Typical results for the miild steel specimens are shown in figure 8.266 D. the hardness is almost independent of the load. and Vickers hardness measurements made at small loads (to give very small impressions) in the free surface of the specimen. the mean pressure on the indenter is almost independent of the size of the indentation. In this way the elastic limit of the deformed material in the free surface around the indentation and in the indentation were determined. If.e. as in the Vicker's test. i. 60 60 401 . therefore./e -i10 w20 + 30 61 10 v 0 01 02 03 04 05 d/D FIGURE 7 06 07 08 09 1. (> : bb30 A 30 0 995 mm. The Vickers hardness numbers were also determined at each stage. Brinell impressions of various sizes were then made in the surface of mild steel and annealed copper specimens. It is seen that the elastic limit of the metal gradually rises as the edge of the indentation is approached.ball diami ~317 u 2. In this case.0 Relations of this type were determined for specimens of mild steel and annealed copper. Tabor pyramidal indenter possessing a large apex angle.

210 FIGunE8. suggest. curve iii = 0 49 and curve iv = 0 23. 0-0 *1 220 230 x .at the edge of the indentationwith the mean pressure used in formingthe indentation. FIGURE Values of d/D in curve i = 0*84.. therefore.60 40- 2 400.. It would.A simple theory of hardness 267 approached..v iii 20 -2 -1 0 I 2 3 distance from edge of indentation (mm. Empiricaltests 80 ~60 30 .appeardifficultto assign a 'representative' value to the elastic limit of the whole material. Thereis also a variationin elasticlimit at variousdepthsin the bulk of the material(O'NeillI934). 80 - ..x Deformation..Results Pm .I | 1 I .curve ii = 0 69.we may compare elastic limit Y.For example.4E 50 40 0 40 20 2200 tI0 O ~~~~~~I 150 160 170 180 190 200 Vickers hardness Elastic limit.however.that the elasticlimit at the edgeof the indentation may be used value for the whole of the deformedmaterialaroundthe as a ? representative' the impression.) 4 9.

268 D. annealed copper: 0 x hardness measurements. mild steel.1 0. . Curve A./mm./mm. Tabor TABLE 9 size of impression d/D 0-27 0-37 0-5 0-23 0-49 0-69 0-84 percentage deformation corresponding to Ye 5 8 9 6 9 15 20 metal annealed copper mild steel Ye (kg. --stress-strain curve.2 03 0 4 dID 05 FiGuRE 10.2) ratio Pm!Y 27 39 44 132 159 161 190 2-6 2-8 2-8 2-6 2-7 2-6 27 2 00 100 - 0 xo~~~~~ B 5 equivalent deformation % 10 15 06 07 08 09 0. Curve B.2) 10-5 14 16 51 57 63 70 Pm (kg.

* Similar results are obtained for metals subjected to deformation under tension. of this table also shows that the deformation corresponding to Yeis approximately proportional to the ratio d/D. Similar measurements were carried out on copper and mild steel specimens that had been compressed by various amounts. Figure 10 shows the results obtained for annealed copper and mild steel. There is close agreement between the two curves. i. assume a more general relation and write (17) a q(=d) where q and y are constants. where c had essentially the same value as before. The hardness values for copper are for annealed specimens that have been deformed in compression experiments by 0. Now compare the stress-strain characteristics with the hardness curves. It was again found that Pm = cY.e. curve A (author). if the deformation is expressed as a percentage 20d/D. It is again seen that there is reasonably close agreement between the hardness curves and the stress-strain curves. We may. the elastic limit has been multiplied by a constant factor and plotted against the deformation or strain. at the edge of the indentation the deformation may be written approximately as S= S. The results given by Kurth (I908) for tensile experiments on drastically annealed copper may be plotted to give a series of curves very similar to those given in figure 11. i.A simple theory of hardness 269 for copper and steel are given in table 9. It is seen that over a wide range of indenta= CYe. The last column tion sizes. 9-6. however.. are given in figure 11. the values of Pmhave been plotted against the values of d/D. 29-6 and 41b5 %. Results for annealed copper. For the stressstrain curves. As we have seen.e.d/D curves that have been displaced along the deformation axis by amounts equal to the initial deformation of the specimen. One then obtains a series of Pm. 1741. 114. The hardness values for mild steel are for specimens that have been deformed in compression experiments by 0. curve B (author). i. the deformation S1 at the edge of the deformation is directly proportional to d/D. the elastic limit is a simple power function of the deformation or strain. Over an appreciable range of deformation. and ordinary bright mild steel. It was also found that the representative deformation was approximately additive to the initial deformation. . to a first approximation. y = b6x (16) where b and x are constants (Nadai I93I).e.* Derivation of Meyer's laws Meyer's laws may readily be derived from equation (15). For the hardness curves. 22 1 and 35 7 %. The analysis may be extended to hardness measurements which have been carried out on specimens that have been deformed by various amounts.+?20d/D.where c has a value lying between 2 6 and 2 8.

this yields Dn-2 (19) Thus for indentations made with balls of different diameters D1. --.5 0 10 20 30 40 50 60 deformation % FIGURE11.k2. (21) . Curve A. = kldn = k2dn = k3dw. Curve B. mild steel.stress-strain.. 28Y 200 - A A 2.. Tabor = = O0 + D)) (18) For completely annealed materials 80 = 0. D3. D2.. annealed copper.11-4 35j7 B ~~10 00 9-6 304Q506 29-6 41. F where k1. = kn3D (20) are given by A = kjDn2 = 3-2. hardness. beq constant and z = xy = constant.. so that where A 17T.1k . = If for convenience we write n z + 2..270 Equation (12) then becomes = D.

the power y in equation (17) is approximately unity.08 0. It is interesting to note that according to the experimental measurements described above.42 0 07 0.25 2-28 (Schwarz n-values) 2-61 2-17 2*40 2-12 2-50 2-14 2.17 0 40 0412 0 50 0-14 0'20 Vikers hardness One may expect that similar considerations will apply in the case of pyramidal and conical indenters. Most of the values given by O'Neill. FROM DATA BY O'NEILL (I944) tensile index x 0. as table 10 shows. -8 . TABLE 10.20 0. so that n in equations (19). For these indenters the indentations are geometrically similar whatever the size of the indentation.A simple theory of hardness 271 The two laws expressed by equations (19) and (20) (see equations (2) and (3)).25 2-44 2-10 2 45 2-25 2.12 0.25 0 07 0-10 metal Norris data: mild steel A yellow brass yellow cold drawn copper L Stead data: steel 1A steel 2A steel 4A steel 6A Schwarz data: brass: annealed rolled annealed copper: rolled nickel: annealed rolled aluninium: annealed typical values of Meyer index n (O'Neill) 2. the 'representative' deformation produced by the indentation and the 'representative' value of the elastic limit will be constant.259 0 404 0494 0. As a result.38 0 04 0-43 0-07 0 15 Kokado-Neill theory: (n -2) 0.414 0-24 0 22 0-19 0-18 0.31 0. and the hardness number H is expressed as the ratio of the applied load to the superficial Vol.25 2. COMPARISON OF MEYER INDEX n AND STRESS-STRAIN.06 0.20 n-2 0-25 0-44 0 10 045 0-25 0-25 025 0-28 0-61 0. were first deduced empirically by Meyer and are found to hold over a wide range of experimental conditions. i92. Consequently. O'Neill (I944) has suggested that n = 2 + 2x. however. the mean pressure Pm required to produce plastic flow should be independent of the size of the indentation if the friction between the indenter and the metal is negligible.12 0-14 0.22 005 0. INDEX x. In the Vickers hardness measurements a shallow pyramidal indenter is used. (20) and (21) is roughly equal to (2 + x). Similar relations are also found to be valid for materials which have been cold-worked by various amounts.23 0. Experiments show that this is generally true.12 0412 0. On the basis of earlier work by Kokado (I925). A. are considerably nearer the relation n =2 + x.

Typical values obtained for the steel and copper specimens used in the earlier experiments are given in table 11. It is also evident from the values of 81 and c that the Vickers hardness numbers will be close to the Brinell hardness numbers over an appreciable range of hardness values. Tabor area of the indentation formed.272 D. This value is assumed to correspond to the 'representative' value of the elastic limit around the indentation. It is compared. (I1945) on the penetration of highly worked copper by hard steel cones. The elastic limit corresponding to a deformation of (60+ 8) %0was then determined from curves similar to those given in figure 9. it is found that H = 0-93Pm Empirical measurements similar to those described above suggest that the 'representative' deformation 61 produced by the deformation corresponds to a value of about 8 to 10 0. the Vickers hardness number is 2-9 to 3 times the 'representative' value of the elastic limit of the material around the indentation. with the observed Vickers hardness numbers.. in table 11.in.in. From the geometry of the indentation. The agreement is reasonably good over a wide range of deformations. whilst the elastic limit of the material was 17-5 tons/sq.5 17'5 25 8= (8+?8) (%) 8 14 18 21 33 8 14 20. For a poorly lubricated cone possessing a semi-angle of 60? (this is approximately the same as that of the Vickers diamond).2) 55 62 66 67 73 15 20 23. Comparison between the stress-strain curve and the * This ratio may be compared with similar experiments described by Bishop et al. The ratio is 3*4. whilst the constant c connecting the mean pressure Pm with the 'representative' value of the elastic limit Y lies between 3 2 and 3 4. .* Consequently. Vickers hardness measurements were made on specimens of these metals after they had been deformed by various amounts 60.5 25-5 33 Discussion Y at 6 (kg./mm. the mean pressure Pm is proportional to the elastic limit Yeof the material at the edge of the indentation. In particular.3 25 26-6 cY 2-9Y 159 176 190 194 211 30Y 45 60 70 75 80 observed Vickers hardness number 156 177 187 193 209 39 58 69 76 81 annealed copper The experimental results recorded in figures 10 and 11 show that there is a welldefined relation between the mean pressure Pm observed in Brinell hardness measurements and the stress-strain characteristics of the metal under consideration. TABLE 11 metal mild steel initial deformation 8o (%) 0 6 10 13 25 0 6 12. the pressure of penetration was 60 tons/sq.

depends on the size of the indentation and is a function of the ratio d/D.cY. is necessarily of a crude nature. By combining these two main results which connect Pm with Y and Y with d/D. This deformation. reasonably close agreement is obtained between the hardness measurements and stress-strain characteristics of various metals. As a result the yield pressure in general increases with the load. It does not take into account the question of friction between the ball and the specimen. which is a dimensionless parameter. where the indentation is geometrically similar whatever its size. In this way a series of hardness measurements with a spherical indenter may be used to determine the degree of workhardening of a given metal. it is possible to explain the well-established empirical laws of Meyer. to a first approximation. In addition. and suggest that the elastic limit at the edge of the indentation provides a mean or representative value for the whole of the deformed material around the indentation. to a rough approximation. first. Council of Scientific and Industrial Research. the work-hardening index x is related to the Meyer index n by the relation x = n -2. for carrying out a large part of the experimental work described in this paper. It is clear from this analysis that hardness measurements are essentially a measure of the elastic limit of the material being examined. Measurements show that the deformation is roughly proportional to d/D and that. and secondly. Australia. a number of well-established empirical relations. the mean pressure to produce plastic flow is almost independent of the size of the indentation. it does describe the main characteristics of hardness measurements for a spherical indenter. Brinson of the Section of Tribophysics. and explains. given in this part of the paper. the amount of work-hardening and hence the elastic limit increases with the size of the indentation. Measurements with spherical indenters thus provide information. The picture of the factors involved in Brinell hardness measurements. In a similar way the theory provides a semi-quantitative relation between the Vickers hardness number and the yield stress of the material. With pyramidal or conical indenters. it is additive to any work-hardening to which the metal may have been subjected in bulk. where c has a value of about 3. the hardness number has a single value over a wide range of loads. Nor does it tell us anything of the detailed way in which each portion of the indentation is plastically deformed and displaced. and the analysis given here shows that. The experiments also show that the elastic limit Yeat the edge of the indentation depends on the amount of deformation produced at this region by the indentation process itself. These results agree with the theoretical results of Hencky and Ishlinsky. about the elastic limit of the material. however. Nevertheless. With spherical indenters. This was first described empirically by Meyer in 1908.A 8imple theory of hardness 273 hardness results show that PI . about the way in which the elastic limit increases with the amount of deformation. Consequently. in a semi-quantitative way. I should also like to thank I8-2 . I wish to express my sincere thanks to Mr G.

Bowden for his constant encouragement. Tabor Professor T. Hill and Dr E. Methodes d'Essai. Instn Civil Engrs.) 8. Phys. Nadai 1931 Plasticity New York. London. 1931 Krupkowski Kurth I908 Z. 8. Lond. des Methodes d'Essai des Materiaux de Construction. 3. Congr. 28. Iron Steel Inst. dtsch. Brinell I90Q II. 445. 115. English translation: Ministry of Supply A. Shore I9I8 J. 59. 9. Amer. 835. 52. 641. 593. 463.S. Hencky 1923 Z. Foss & Brumfield I922 Proc. Camb. Finally. Ishlinsky I944 J. Mag. Int. 107. Tohoku Univ. Engrs 2. No. A. Kokado 1925 J. dtsch. O'Neill 1944 Proc. Oxford University Press. Soc. London. London. 57. 147. (U. and the Ministry of Supply (Air) for a Research Grant. Cherry and Dr W. 486 (James Forrest Lecture). Ver. Batson I9I8 J. Instn Mech. Hill. 43. 1560. Soc. Ing. Engrs. Raman I9I8 Phys. 233. Rev. Okubo I922 O'Neill 1934 The hardness of metals and its measurements. Hill & Mott 1945 Proc. 273. Prescott 1927 Applied elasticity. Rev. 2. Sci. Mech. Timoshenko 1934 Theory of elasticity. 857. 332. Soc. 324. Metall. REFERENCES Andrews Andrews Phil. Paris. Boas (Melbourne). Test. Pro?. M. Phil.274 D. 740. I wish to thank Dr F. 2/47. 645. Roy. 151. Orowan (Cambridge) for valuable discussions. 28. Engrs Japan. Vincent I900 1930 1931 . Martel I895 Comm. 52. Ing. 241. Ver. New York: McGraw Hill. 26. Math. Carpenter & Robertson 1936 Metals. Lee & Tupper 1947 Proc. Proc. Theoretical Research Translation. Mat. Hertz I896 Miscellaneous papers. 12. Bishop. Meyer I908 Z. Soc. Mech. Math. Edwards & Austin 1923 J.D. 22. Iron Steel Inst. angew. Taylor 1946 J. 442. Appl.R. Mr R. 11.S. Mech. Soc.R. P. 312. Phys. Instn Mech. 188. Rep.. Soc. 10.