The Physical and Mechanical Properties of Metals …2 The Physical and Mechanical Properties of Metals and Alloys 2.10 Of well over one hundred elements—if we include the increasing

2The Physical andMechanical Properties ofMetals and Alloys

2.10 Of well over one hundred elements—if we include the increasingnumber of man-made ones—only eighteen have definite non-metallicproperties. Six are usually classed as 'metalloids'—elements like silicon,germanium and arsenic—in which physical and chemical properties aregenerally intermediate between those of metals and non-metals, but theremainder have clearly defined metallic properties. Metals are generallycharacterised by their lustrous, opaque appearance and, in respect of otherphysical properties, metals and non-metals contrast strongly.

As we have seen (1.76) a metal consists of an orderly array of ionssurrounded by and held together by a cloud of electrons. This is reflectedin many of the physical properties of metals.2.11 Melting point All metals (except mercury) are solids at ambienttemperatures and have relatively high melting points (see Table 1.1) whichvary between 234K (-390C) for mercury and 3683K (34100C) for tungsten.Non-metals include gases, a liquid (bromine) and solids. Their meltingpoints vary much more widely: between IK (-2720C) for helium andapproximately 5300K (50000C) for carbon.2.12 Density The relative density (formerly specific gravity) of a materialis defined as

the weight of a given volume of the materialthe weight of an equal volume of water.

Metals generally have higher relative densities (Table 1.1) than non-metals. Values vary between lithium (0.534) which will float in water andosmium (22.5) which is almost twice the density of lead, which suggeststhat the simile 'as heavy as lead' needs revision.

2.13 Electrical conductivity Non-metals are generally very poor con-ductors of electricity, indeed those where the bonding is entirely covalentwill be insulators since all valance electrons are held captive in individualbonds and can move only in restricted orbits. By comparison in metalselectrical conductivity arises from the presence of a sea or cloud of mobileelectrons permeating the static array of ions. The electrons are able to flowthrough the ion framework when a potential difference is applied acrossthe ends of the metal—which may be many miles apart as in the electricgrid system. As indicated in Table 2.1 the electrical and thermal conduc-tivities of metals follow roughly the same order. This is to be expectedsince both the flow of electricity and heat depend upon the ability ofelectrons to move freely within the metallic structure. For purposes ofsimple comparison Table 2.1 relates electrical and thermal conductivitiesof some important metals to those of silver (100). Although silver is mar-ginally superior in terms of electrical conductivity to copper, the latter isused industrially because of relative costs. In fact for power transmissionthrough the national grid aluminium lines are generally used for reasonsgiven later (17.13). Electrical conductivity is reduced by alloying and thepresence of impurities (16.21) as well as by mechanical straining.

Table 2.1 Relative electrical and thermal conductivities of some metals

Metal Relative electrical conductivity Relative thermal conductivity

Silver 100 100Copper 96 94Gold 69.5 70Aluminium 59 57Magnesium 41 40Beryllium 40 40Tungsten 29 39Zinc 27 26.5Cadmium 22 22Nickel (23) 21Iron 16 17Platinum 15 17Tin 12.5 15.5Lead 7.7 8.2Titanium 2.9 4.1Mercury 1.6 2.2

Electrical conductivity is measured in units Sm"1, where the unit ofconductance the Siemen (S), is equivalent to Q "!. Generally it is moreconvenient to consider the electrical resistivity (Q) of a material which isthe inverse of its conductance and is of course measured in flm. Resistivityvaries with temperature and over a limited temperature range a linearrelationship of the form:

Rt = R(I + Qt)

holds good. Here Rt is the resistance at the upper temperature, R theinitial resistance and t the increase in temperature. Q is the temperaturecoefficient of resistance of the material.

2.14 Thermal conductivity This arises in a similar way to electrical con-ductivity. Electrons pick up kinetic energy from the increased vibrationsof the ions where the metal is hot. They pass rapidly through the ionframework where they collide with distant ions, causing them in turn tovibrate more rapidly. In this manner electrons behave as transporters ofenergy.

Metals are very good conductors of heat whereas most non-metals arenot. The flow of heat in a conductor is governed by:

Q=-X fdx

where Q is the heat flow across unit area, X is the coefficient of thermalri T

conductivity and T the temperature. — will be the 'temperature gradient'dx

at that unit area. Thermal conductivity is measured in units, Wm-1K"1.2.15 Specific heat capacity The specific heat capacity (Cp, Cv) of asubstance is the quantity of heat required to raise the temperature of lkgof the substance IK. The units are JlCg-1K"1. The specific heat capacitiesof metals are low compared with those of non-metals so that it is lessexpensive to raise their temperatures.

Dulong and Petit's Law states that for all elements the product of thespecific heat capacity and the atomic weight is approximately constant andthis product is called the Atomic heat. This law was used more than acentury ago to assess the atomic weights of many (then) new elements. Itinvolves the relationship between heat capacity and the vibrational energyof atoms.

Table 2.2 Thermal properties of some important metals

Metal Coefficient of thermal expansion (^) Specific heat capacity(K"1 x 10"6) (J kg"1 K'1)

Aluminium 23 913Copper 17 385Gold 14 132Iron 12 480Lead 29 126Magnesium 25 1034Nickel 13 460Silver 19 235Tin 23 226Titanium 9 523Zinc 31 385

2.16 Thermal expansion As materials are heated the amplitude ofatomic vibrations increases and this is evident as an increase in volume.The coefficient of cubic expansion (y) is the increase in volume per unit

* This is basically similar to Ohm's Law governing the flow of electricity (electrons) through a con-ductor.

volume per unit rise in temperature (unit, K l). Similarly the coefficient oflinear expansion (a) is the increase in length per unit length per unit rise

in temperature (unit, K"1) or a = - -

where lo = original length; lt = length after a rise in temperature of t.By suitable alloying additions to some metals it is possible to reduce a

to low limits. Thus Invar (13.25) is used in long measuring tapes, pendulumrods for observatory clocks (before the days of electronic timekeeping),etc., whilst similar alloys are used in the delicate sliding mechanisms ofinstruments used under conditions of widely varying temperature, eg mili-tary rangefinders used in desert warfare. Further alloys are also used inbimetallic strips in small thermostats where the differential expansion ofthe two alloys of the strip leads to bending of the unit and a make/breakcontact.2.17 Behaviour to light Most metals reflect all wavelengths of lightequally well for which reason they are white or nearly so. Notable excep-tions are copper and gold whilst zinc is very faintly blue and lead slightlypurple.

The reflecting capacity of metals is yet another aspect of the mobility ofits electrons; an incident light wave causes the electrons near the surfaceof a metal to oscillate and as a result the incident wave is reflected backinstead of being absorbed by the metal. Thus the reflection in a mirror isdue to the oscillation caused in silver's mobile electron cloud.2.18 Behaviour to short-wavelength radiations Metals are transparentto y-rays (2.93) and to those X-rays of short wavelength ('hard' X-rays)(2.91).2.19 Magnetic properties Most metals are magnetic to some slightextent but only in the metals iron, nickel, cobalt and gadolinium is magnet-ism strong enough to be of practical interest. The pronounced magnetismof this group is called 'ferromagnetism' (14.30).

Whilst many of these physical properties such as conductivity, magnet-ism and melting point dictate special uses for metals, it is mechanicalproperties such as strength, ductility and toughness which concern us prin-cipally in engineering design.

Fundamental Mechanical Properties

2.20 Whereas the directional nature of the covalent bond results in theextreme rigidity of substances like diamond and quartz, the non-directionalnature of the metallic bond makes it relatively easy to bend a piece ofmetal. Moving groups of metallic ions through the electron 'sea' can beachieved in a number of ways such as hammering, rolling, stretching andbending. Fundamental mechanical properties of metals are related to the

amounts of deformation which metals can withstand under different cir-cumstances of force application. Ductility refers to the capacity of a sub-stance to undergo deformation under tension without rupture, as in wire-or tube-drawing operations. Malleability, on the other hand, is the capacityof a substance to withstand deformation under compression without rup-ture, as in forging or rolling. Substances which are highly ductile are alsohighly malleable but the reverse may not be true since some extremelymalleable substances are weak in tension and therefore liable to tear.Moreover, whilst malleability is usually increased by raising the tempera-ture (for which reason metals and alloys are often hot-forged or hot-rolled),ductility is generally reduced by heating, since strength is also reduced.

2.21 Toughness refers to a metal's ability to withstand bending or theapplication of shear stresses without fracture. Hence, copper is extremelytough, whilst cast iron is not. Toughness should not, therefore, be confusedwith either strength or hardness, properties which will be discussed later.

2.22 Since these fundamental mechanical properties of ductility, malle-ability and toughness cannot be expressed in simple quantitative terms, ithas become necessary to introduce certain mechanical tests which arerelated to these properties and which will allow of comparative numericalinterpretation. Moreover, the engineer is more concerned with the forceswhich cause deformation in metals rather than with the deformation itself.Consequently tensile tests and hardness tests correlate the amounts ofdeformation produced with given forces in tension and compressionrespectively, whilst impact tests are an almost direct measurement oftoughness. Such precise measurements of force-deformation values makeit possible to draw up sets of specifications upon which the mechanicalengineer can base his design.

Tenacity or Tensile Strength

2.30 The tensile strength of a material is defined as the maximum forcerequired to fracture in tension a bar of unit cross-sectional area. In practicea test-piece of known cross-sectional area is gripped in the jaws of a testingmachine and subjected to a tensile force which is increased by suitableincrements. For each increment of force the amount by which the lengthof a pre-determined 'gauge length' on the test piece increases is measuredby some device. The test piece is extended in this way to destruction.

A force-extension diagram can then be plotted (Fig. 2.1). At first the rateof extension is very small and such extension as there is is directly pro-portional to the applied force; that is, OQ is a straight line. If the appliedforce is removed at any point before Q is reached the gauge-length will returnto its original dimensions. Thus the extension between O and Q is elasticand the material obeys Hooke's Law, which states that, for an elasticbody, the strain produced is proportional to the stress applied. The valueStress

— is constant and is equivalent to the slope of OQ. This constant valueStrain

Fig. 2.1 The force-extension diagram for an annealed low-carbon steel.

is known as Young's Modulus of Elasticity (E) for the material. Considera test piece of original length, L, and cross-sectional area, 'a', stretchedelastically by an amount, T, under a force, P, acting along the axis of thespecimen, then:

Longitudinal StressLongitudinal Strain

= P/al/L

_ P L~~aT

Young's Modulus is in fact a measure of the stiffness of the material intension. This value and the stress range over which it applies are of greatimportance to the engineer. Young's modulus is measured in the sameunits as those of stress, since:

_ StressE = —

Strain

[Stress]~ [length]/[length]

= [Stress]

FO

RC

E (

P)

elasticextension

uniform plastic extension

EXTENSION [Jt)

'necking

If at any point on the part of the curve under consideration the force isrelaxed then the test piece will return to its original length, extension sofar being entirely elastic.

2.31 If the test piece is stressed beyond the point O the curve deviatesfrom its straight-line characteristics. Q is therefore known as the elasticlimit or limit of proportionality and if the force is increased beyond thispoint a stage is reached where a sudden extension takes place for noincrease in the applied force (assuming that we are testing a specimen ofannealed low-carbon steel as indicated in Fig. 2.1). An explanation of thisphenomenon, known as the yield point, R, will be given later (8.61). Ifthe force is now removed the elastic extension will disappear but a smallpermanent plastic extension or permanent set will remain.

As the force is increased beyond the point R the test piece stretchesrapidly—first uniformly along its entire length and then locally to form a'neck'. This 'necking' occurs just after the maximum force value has beenreached at S, and since the cross-section decreases rapidly at the neck, theforce at B required to break the test piece is less than the maximum forceapplied at S.

This might be an appropriate moment at which to mention the differencebetween a 'force/extension' diagram and a 'stress/strain' diagram sincethese terms are often loosely used by both metallurgists and engineers.Fig. 2.1 clearly represents a force/extension diagram since total force isplotted against total extension, and, as the force decreases past the pointS, for reasons just mentioned, the decrease is indicated on the diagram.Stress however is measured as force per unit area of cross-section of thetest-piece and if we wished to plot this we would need to measure theminimum diameter of the test piece at each increment of applied force.This would be particularly important for values of force after the point S,since from S onwards the effective cross-section is decreasing rapidly dueto the formation of the neck. The test piece is only as strong as the forceits minimum diameter will support.

If stress were calculated on this decreasing cross-section the resultingstress/strain diagram would follow a path indicated by the broken line toBi from S onwards. In practice, however, a nominal value of the tensilestrength of a material is calculated using the maximum force (at S) and theoriginal cross-sectional area of the test piece. Therefore:

Maximum force usedTensile strength = _ . .— :—Original area or cross-section

In this connection the term 'engineering stress' is often used; it implies theforce at any stage of the loading cycle divided by the original area ofcross-section of the material.

Although tensile strength is a useful guide to the mechanical propertiesof a material it is not of paramount importance in engineering design.After all, the engineer is not particularly interested in a material once itbegins to stretch plastically—unless of course he is a production engineerengaged in deep-drawing or some other metal-forming process. In the case

of structural or constructional engineering, the elastic limit, Q, will be offar greater significance than tensile strength.

2.32 The form of force/extension diagram described above is in fact aspecial case, obtained only for wrought irons and low-carbon steels in thesoft condition (8.61). Most alloys, particularly if they have been heat-treated or cold-worked, show neither a definite elastic limit nor a yieldpoint and give, on test, diagrams of the types shown in Figs. 2.2 and 2.3.

FO

RC

EFO

RC

E

EXTENSION

Fig. 2.2 The effects of heat-treatment on the force-extension diagram of carbon steel. (A)is in the quenched condition; (B) is quenched and tempered; and (C) represents the annealedcondition.

EXTENSION

Fig. 2.3 Typical force-extension diagrams for a non-ferrous alloy, showing the absence ofa well-defined yield point. (D) represents the cold-worked condition, and (E) the fully annealedcondition.

Since the yield point is of greater importance to the engineer than thetensile strength itself, it becomes necessary to specify a stress which corre-sponds to a definite amount of permanent extension as a substitute for theyield point. This is commonly called the 'Proof Stress', and is derived asshown in Fig. 2.4. A line BC is drawn parallel to the line of proportionality,from a pre-determined point B. The stress corresponding to C will be theproof stress—in the case illustrated it will be known as the '0.1% proofstress', since AB has been made equal to 0.1% of the gauge length. Thematerial will fulfil the specification therefore if, after the proof force isapplied for fifteen seconds and removed, a permanent set of not more than0.1% of the gauge length has been produced. Proof lengths are commonly0.1 and 0.2% of the gauge length depending upon the type of alloy. Thetime limit of 15 seconds is specified in order to allow sufficient time forextension to be complete under the proof force.

Fig. 2.4 Method used to obtain the 0.1 % proof stress.

2.33 In addition to determining the tensile strength and the proof stress(or, alternatively, the yield stress), the percentage elongation of the testpiece at fracture is also derived. This is an almost direct measure of duc-tility. The two ends of the broken test piece are fitted together (Fig. 2.5)so that the total extension can be measured.

In order that values of percentage elongation derived from test pieces ofdifferent diameter shall be comparable, test pieces should be geometricallysimilar, that is, there must be a standard relationship or ratio betweencross-sectional area and gauge length. Test pieces which are geometricallysimilar and fulfil these requirements are known as proportional test pieces.They are commonly circular in cross-section. BSI lays down that, for pro-portional test pieces:

FO

RC

E

PROOF FORCE

O l ° /o OFGAUGE LENGTH

Fig. 2.5 The determination of percentage elongation and percentage reduction in area.

where L0 is the gauge length and So the original area of cross-section. Thisformula has been accepted by international agreement and SI units areused. For test pieces of circular cross-section it gives a value

L0 ^ 5d

where 'd' is the diameter at the gauge length. Thus a test piece 200 mm2

in cross sectional area will have a diameter of 15.96 mm (16 mm) andhence a gauge length of 80 mm. Some old tensile testing machines maystill be calibrated in 'tons force'. Since 10 kN = 1.00361 tonf, dual valuescales are not necessary, since, within the accuracy required, 1 tonf = 10kN.

2.34 The smallest diameter, S0, of the neck is measured and from itthe percentage reduction in area calculated (Fig. 2.5). Thus, from ourcomplete set of observations we can derive the following values:

, , , . , . , . Yield force(a) Yield stress = . . : —

Original area of cross-section/ Proof force \or Proof stress = :\ Original area of cross-section J

Maximum force(b) Tensile strength = ——— -c :—

Original area of cross-sectionIncrease in gauge length x 100

(c) Percentage elongation = — ——Original gauge length

(d) Percentage reduction in area =

(Original area of cross-section — Final area of cross section) x 100Original area of cross-section

In terms of SI units stress is measured in N/m2. However, since it isdifficult to appreciate the very large force necessary to break a test piece

S0 (original area of cross section)

'neck'

gaugelength, L0

°/o Elongation

°/o Reductionin area

Plate 2.1 The Avery-Denison Servo-controlled Tensile Testing Machine, with an appliedforce capacity of 60OkN.

The straining unit which is shown on the left embodies a double-acting hydraulic cylinderand ram. The force on the test piece is measured by load-sensitive cells and is indicated onthe display panel of the control console shown on the right. The full load/extension cycle iselectronically controlled and a permanent trace is produced. (Courtesy of Messrs Avery-Denison Ltd, Leeds).

of one square metre in cross-section, most bodies, including BSI, quotetensile stress in metals in N/mm2. This at least enables the student to relatethe tensile strength of a steel to the force necessary to break in tensionone of the thicker steel strings on his guitar.

2.35 Early tensile-testing machines were of the simple beam type inwhich the applied force was magnified by using a first-order lever system.With such machines an accurate evaluation of extension was possible up tothe elastic limit by using a sensitive extensometer but beyond the maximumforce, determinations of force/extension characteristics were impossiblebecause the test piece fractured quickly as soon as necking began, therebeing no means of relaxing the applied force rapidly enough. Modernmachines however are usually servo-hydraulically loaded (Plate 2.1) so thata complete force/extension relationship can be obtained. Since advancedcomputer control technology is now employed automatic calculation ofproof stress, yield stress, ultimate tensile stress and percentage elongationare carried out; whilst software is available for cycling and data storage.Software programs can be written to meet other specific requirements.These machines can also be used for compression and transverse testing,and vary in size between large machines with a capacity of 1300 kN andsmall bench models having a capacity of only 20 kN.

2.36. In situations where a large amount of energy is being expendedagainst gravity as in various types of aero-space travel—or even drivingthe humble 'tin Lizzie' up a hill—it becomes necessary to relate the tensileproperties of a material to its relative density. Thus, what used to be calledthe 'strength-to-weight ratio' became important in the design of both landand air transport vehicles. In modern terminology this became 'specificstrength'. Thus:

.„ , Tensile strength of materialSpecific strength = —

Relative density of materialWhen stiffness is the prime consideration, however, Young's Modulus ofElasticity is a more appropriate guide to the required properties and avalue termed specific modulus is now generally accepted as being relevant,ie—

Young's modulus of elasticitySpecific modulus of elasticity = iTTT-—"A %

Relative density

Hardness Tests

2.40 Classically, hardness could be defined as the resistance of a surfaceto abrasion, and early attempts to measure surface hardness were basedon this concept. Thus in the Turner Sclerometer a loaded diamond pointwas drawn across the surface of the test piece and the load increased untila visible scratch was produced. In Moh's Scale—still used to evaluate thehardness of minerals—substances were arranged in order of hardness such

that any material in the scale would scratch any material listed below it.Thus diamond (with a hardness index of 10) heads the list whilst talc (withan index of 1) is at the foot of the scale.

Whilst such methods undoubtedly reflect a true concept of the funda-mental meaning of hardness, they have been abandoned in favour ofmethods which are capable of greater accuracy but in which the resistanceof the surface layers to plastic deformation under static pressure ismeasured rather than true hardness. In most of these methods the staticforce used is divided by the numerical value of the surface area of theresulting impression to give the hardness index.

2.41 The Brinell Test, probably the best known of the hardness tests,was devised by a Swede, Dr. Johan August Brinell in 1900. In this test ahardened steel ball is pressed into the surface of the test piece using theappropriate specified force. The diameter of the impression so producedis then measured and the Brinell Hardness Number, HBy derived from:

Force, PB Surface area of impression

It can be shown that the surface area of the impression isn—l D - yJD2 - d2 j where D is the diameter of the ball and 'd' the diameter

of the impression (Fig. 2.6). Since we are dealing with the actual area ofthe curved surface of the impression the derivation of the above expressionis quite involved.Hence,

and the units will be kgf/mm2. To obviate tedious calculations HB is foundby reference to the appropriate set of tables.

2.42 It is obviously important that the stress produced by the indenterat the surface of the test piece shall suit the material being tested. If forexample in testing a soft metal we use a force which is too great relativeto the diameter of the ball, we shall get an impression similar to that

Fig. 2.6 The relationships between ball diameter, depth of impression and dimensions ofthe test piece in the Brinell-type test.

not lessthan 8h

TEST PIECE

not lessthan 3D

indicated in Fig. 2.7A. Here the ball has sunk to its full diameter and theresult is obviously meaningless. The impression shown in Fig. 2.7B onthe other hand would be obtained if the force were too small relative tothe ball diameter and here the result would be likely to be very uncertain.For different materials then, the ratio P/D2 has been standardised in orderto obtain accurate and comparable results. P is measured in kgf and D inmm.

Material Approximate HB range PID2 ratio used

Steel and cast iron Over 100 30Copper, copper alloys and .~aluminium alloys 30-200Aluminium 15-100 5Tin, lead and their alloys 3-20 1

INCORRECT C O R R E C T

Fig. 2.7 The influence of depth of impression on the accuracy of a Brinell determination.

Thus in testing a piece of steel we can use either a 10-mm ball in conjunc-tion with a 3000 kgf load; a 5 mm ball with a 750 kgf load; or a 1 mm ballwith a 30 kgf load. In the interests of accuracy it is always advisable to usethe largest ball diameter that is possible. The limiting factors will be thewidth and thickness of the test piece, and the small ball would be used forthin specimens, since by using the large ball we would probably be, ineffect, measuring the hardness of the table supporting our test piece. Thethickness of the specimen should be at least eight times the depth, 'h', ofthe impression (Fig. 2.6). Similarly the width of the test piece must beadequate to support the applied force and it is recommended that thedistance of the centre of the indentation from the edge of the test pieceshall be at least three times the diameter of the indenting ball.

2.43 The Vickers Hardness Test—or Diamond Pyramid HardnessTest —uses as its indenter a diamond square-based pyramid (Fig. 2.8)which will give geometrically similar impressions under different appliedforces.This eliminates the necessity of deciding the correct P/D2 ratio as isrequired in the Brinell test. Moreover, the diamond is more reliable forhard materials which have a hardness index of more than 500, since it doesnot deform under pressure to the same extent as a steel ball. Using thediamond point, however, does not eliminate the necessity of ensuring that

A B C

Fig. 2.8 The Diamond Pyramid lndentor and its resulting impression.

the thickness of the specimen is sufficient, relative to the depth of theimpression.

In this test the diagonal length of the square impression is measured bymeans of a microscope which has a variable slit built into the eyepiece(Fig.2.8 (iii)). The width of the slit is adjusted so that its edges coincidewith the corners of the impression and the relative diagonal length of theimpression then obtained from a small instrument attached to the slit whichworks on the principle of a revolution counter. The ocular reading thusobtained is converted to Vickers Pyramid Hardness Number by referenceto tables. The hardness index is related to the size of the impression in thesame way as is the Brinell number.

Fig. 2.9 Basic principles of the force application system in the Vickers Hardness TestingMachine.

2.44 The Rockwell Test was devised in the USA, and is particularlysuitable for rapid routine testing of finished material since it indicates thefinal result direct on a dial which is calibrated with a series of scales. Anumber of different combinations of indenter and indenting force can beused in conjunction with the appropriate scale:

fulcrum rigid beam

leadcounterpoise

diamond .pyramid

testpiece

to automatictiming mechanism

at-rest'support

load

variableslit

coupled toocular counter

impression(iii)

(i)

(i»)

Scale Indenter Total force (kgf)A Diamond cone 60B 1 V steel ball 100C Diamond cone 150D Diamond cone 100E Vs" steel ball 100F Vi6" steel ball 60G V16" steel ball 150H Vs" steel ball 60K W steel ball 150

Of these, scale C is probably the most popular for use with steels.

Fig. 2.10 The Rockwell Diamond Cone lndentor.

2.45 The Shore Scleroscope (Greek: 'skleros'—hard) tests thematerial very near to its surface. The instrument embodies a small dia-mond-tipped 'tup' which is allowed to fall from a standard height inside agraduated glass tube. The height of rebound is taken as the hardness index.Since the Shore Scleroscope is a small, portable instrument, it is very usefulfor the determination of hardness of large rolls, castings and gears, andother large components which could not easily be placed on the testingtables of any of the more orthodox testing machines.

The development of digital display units has enabled very small portablehardness testers of the indentation type to be manufactured. One of theseconsists of a small motorised probe which, when pressed against the surfaceof the test piece, makes a minute diamond impression using a force ofonly 8.4 N. Consequently such a test is virtually non-destructive and theinstrument can be used in the most remote corners of the factory, hangaror repair yard. At the same time a high accuracy of ± 15 VPN is claimedover the hardness range of 50 to 995 VPN. Such instruments have largelyreplaced the Shore Scleroscope in terms both of accuracy and adaptability.

Table 2.3 gives representative hardness numbers, together with othermechanical properties, for some of the better-known metals and alloys.

O»2 mmradius

Table 2.3 Typical mechanical properties of some metals and alloys

Metal or Alloy

Lead

Aluminium

Duralumin

Magnesium-6AI/1Zn

Copper

70/30 Brass

Phosphor bronze(5% tin)

Mild steel

0.45% carbonsteel

4Ni/Cr/Mo steel

18/8 stainlesssteel

Grey cast iron

Titanium(commerciallypure)

Titanium alloy(4Sn/4AI/4Mo0.5Si)

Condition

Soft sheet

Wrought andannealed

Extruded andfullyheat-treated

Extruded bar

Wrought andannealed

Annealed

Deep-drawn

j Rolled andannealed

Hard-rolled

Hot-rolledsheet

Normalised

0.7%ProofStrength(N/mm2)

275

170

46

85

370

120

650

270

420

Water-quenchedand temperedat 6000C

Air-hardenedandtempered at3000C

Softened

As cast

Annealedsheet

Precipitationhardened

540

1200

185

370

1200

TensileStrength(N/mm2)

18

60

430

300

216

320

465

340

710

400

665

780

1550

525

300

450

1390

SpecificStrength(N/mm2)

1.54

21.8

154

167

24.1

37.6

54.6

38.1

79.6

50.8

84.7

99.4

198

66.3

40.5

100

309

Young'sModulus(kN/mm2)

16

70

71

48

130

100

101

210

oc\c\

225

220

150

120

150

SpecificModulus(kN/mm2)

1.37

25.9

25.4

26.7

14.5

11.7

11.3

26.7

QC A

28.7

27.8

20.3

26.7

33.3

Elongation(%)

65

60

15

10

60

68

19

66

5

28

27

25

12

30

0

30

16

Hardness(Brinell)

4

15

115

60

42

62

132

72

188

100

152

200

444

170

250

ImpactValue(IZOd)(J)

27

22

8

59

90

61

75

44

65

22

68

1

61

Bold type denotes maximum value in that property (where relevant).

Fig. 2.11 The Avery-Denison Universal impact-testing machine.This machine can be used for either Charpy or Izod impact tests. For Izod tests, the

pendulum is released from the lower position, to give a striking energy of 170 J; and for theCharpy test it is released from the upper position, to give a striking energy of 300 J. (Thescale carries a set of graduations for each test.) The machine can also be used forimpact-tension tests.

Impact Tests

2.50 Impact tests indicate the behaviour of a material under conditionsof mechanical shock and to some extent measure its toughness. Brittleness—and consequent lack of reliability—resulting from incorrect heat-treatment (13.42) or other causes may not be revealed during a tensile testbut will usually be evident in an impact test.

2.51 The Izod Impact Test In this test a standard notched specimenis held in a vice and a heavy pendulum, mounted on ball bearings, isallowed to strike the specimen after swinging from a fixed height. Thestriking energy of 167 J (120 ft lbf) is partially absorbed in breaking thespecimen and, as the pendulum swings past, it carries a pointer to itshighest point of swing, thus indicating the amount of energy used in fractur-ing the test piece.

TEST-PIECECLAMPINGLEVER

CHUTE FOR BROKENTEST-PIECES

PENDULUMRELEASELEVER IZOO TEST-

PIECE

PENOULUM REST STOP

SECTION THROUGHPENDULUM BO8

AT X - Y

IZOD RELEASEPOSITION

PENDULUM

STRIKER

CHARPY RELEASE POSITION

SCALE

2.52 The Charpy Test, developed originally on the Continent but nowgaining favour in Britain, employs a test piece mounted as a simply-supported beam instead of in the cantilever form used in the Izod test (Fig.2.12). The striking energy is 300 J (220 ft lbf).

2.53 To set up stress concentrations which ensure that fracture doesoccur, test pieces are notched. It is essential that notches always be stan-dard, for which reason a standard gauge is used to test the dimensionalaccuracy of the notch. Fig. 2.12 shows standard notched test pieces forboth the Izod and Charpy impact tests.

Fig. 2.12 Dimensions of standard test pieces for both Izod and Charpy tests.In the Izod test piece, notches 28mm apart may be cut in three different faces so that a morerepresentative value is obtained.

2.54 The results obtained from impact tests are not always easy tointerpret, and some metals which are ductile under steady loads behave asbrittle materials in an impact test. As mentioned above, however, theimpact test gives a good indication of how reliable the material is likely tobe under conditions of mechanical shock. These tests are most likely to bespecified for constructional steels of medium-carbon content.

Other Destructive Tests

2.60 These are often designed specifically for the measurement of someproperty peculiar to a single class of material or to assess the suitability ofa material for a special purpose.

2.61 The Erichsen Cupping Test (11.54-Pt. II) is closely connectedwith the ductility of a material but is in fact designed to assess its deep-drawing properties.

2.62 Compression Tests are used to measure the capability of a castiron to carry compressive loads. A standard test cylinder is tested in com-pression, usually employing a tensile testing machine running in 'reverse'.

2.63 Torsion Tests of various types are sometimes applied tomaterials in wire and rod form.

IOmm SQUARE

HAMMER

ROOTRADIUS

HAMMER

V I C E

SQUARE

IZOD

DETAILS OFNOTCH FORBOTH TESTS CHARPY

Non-destructive Tests

2.70 The mechanical tests already mentioned are of a destructive natureand are subject to the availability of separate test pieces which are reason-ably representative of the production material. Thus, wrought productssuch as rolled strip, extruded rod and drawn wire are generally uniform inmechanical properties throughout a large batch and can be sampled withconfidence. Parts which are produced individually, however, such as cast-ings and welded joints, may vary widely in quality purely because they aremade individually and under the influence of many variable factors. If thequality of such components is important and the expense justified—as inthe case of aircraft castings—it may be necessary to test each componentindividually by some form of non-destructive test. Such tests seek to detectfaults and flaws either at the surface or below it, and a number of suitablemethods is available in each case.

The Detection of Surface Faults

2.80 It is often possible to detect and evaluate surface faults by simplevisual examination with or without the use of a hand magnifier. The pres-ence of fine hair-line cracks is less easy to detect by visual means and someaid is generally necessary. Such surface cracks may be associated withthe heat-treatment of steel or, in a welded joint, with contraction duringcooling.

2.81 Penetrant Methods In these methods the surface to be examinedis cleaned and then dried. A penetrant fluid is then sprayed or swabbedon the surface which should be warmed to about 900C. After sufficienttime has elapsed for the penetrant to fill any fissures which may be presentthe excess is flushed from the surface with warm water (the surface tensionof the water is too high to allow it to enter the narrow fissure). The testsurface is then carefully dried, coated with fine powdered chalk and setaside for some time. As the coated surface cools, it contracts and penetranttends to be squeezed out of any cracks, so that the chalk layer becomesstained, thus revealing the presence of the cracks. Most penetrants of thistype contain a scarlet dye which renders the stain immediately noticeable.Aluminium alloy castings are often examined in this way.

Penetrants containing a compound which fluoresces under the action ofultra-violet light may also be used. This renders the use of messy chalkunnecessary. When the prepared surface is illuminated by ultra-violet light,the cracks containing the penetrant are revealed as bright lines on a darkbackground. Penetrant methods are particularly useful for the examinationof non-ferrous metals and austenitic (non-magnetic) steels.

2.82 Magnetic Dust Methods consist in laying the steel componentacross the arms of a magnetising machine and then sprinkling it with aspecial magnetic powder. The excess powder is blown away, and any cracksor defects are then revealed by a bunch of powder sticking to the area on

Fig. 2.13 The penetrant method of crack detection.(i) The cleaned surface is coated with penetrant which seeps into any cracks present, (ii)Excess penetrant is removed from the surface, (iii) The surface is coated with chalk. As themetallic surface cools and contracts, penetrant is expelled from the crack to stain the chalk.

each side of the crack. Since the crack lies across the magnetic field, lines offorce will become widely separated at the air gap (Fig. 2.14) and magneticparticles will align themselves along the lines of force.

Fig. 2.14 The principles of magnetic particle crack detection.

The Detection of Internal Defects

2.90 Internal cavities in the form of blow holes or shrinkage porosity maybe present in castings of all types, whilst wrought materials may containslag inclusion and other subcutaneous flaws. Welded joints, by the natureof their production methods, may contain any of these defects. Since metalsare opaque to light, other forms of electromagnetic radiation of shorterwavelength (X-rays and y-rays) must be used to penetrate metals and soreveal such internal discontinuities. Although the railway wheeltapper was,for some obscure reason, always 'good for a laugh' at the mercy of theprofessional comedian, he was in fact using a 'sonic' method of testing thecontinuity of structure of the wheel and modern sophisticated methods ofultra-sonic testing use similar principles.

2.91 X-ray Methods X-rays are used widely in metallurgical researchin order to investigate the nature of crystal structures in metals and alloys.Their use is not confined to the research laboratory, however, and manyfirms use X-rays in much the same way as they are used in medical radiogra-phy, that is, for the detection of cavities, flaws and other discontinuities incastings, welded joints and the like.

penetrant

fissure

(i) (H) (Hi)

chalk

component magnetic field

N S

X-rays are produced when a stream of high-velocity electrons impingeson a metal target. Fig. 2.15 illustrates the principle of an X-ray tube inwhich a filament supplies free electrons. Being negatively charged theseelectrons are accelerated away from the cathode towards the anode (some-times called the 'anti-cathode') by the high potential difference betweenthe electrodes. Collision with the anode produces X-rays. The containingtube is under vacuum, as the presence of gas molecules would obstruct thepassage of relatively small electrons. Nevertheless only about 1% of theenergy expended produces X-rays the remainder being converted to heat.Consequently the anode must be water-cooled. For greater output ofX-rays (above 1 MeV) other types of generator such as the 'linear gener-ator' or 'Linac' have to be used.

Fig. 2.15 Radiography of a casting using X-rays.

The penetrating power of electromagnetic radiation generally, dependsupon its frequency. Thus radiation at the ultra-violet end of the visiblespectrum will penetrate our skin to a depth of less than 1 mm but willnevertheless produce painful radiation burns (and possibly skin cancer) ifwe sunbathe carelessly. X-rays, having a much higher frequency than UVlight, will penetrate more deeply and the 'harder' the X-rays (ie the higherthe frequency) the greater the depth of penetration. X-rays used in metal-lurgical radiography are harder than those used in medicine, and are betterable to penetrate metals. At the same time their properties make themmuch more dangerous to human body tissue, and plant producing radiationof this type needs to be carefully shielded in order to prevent the escapeof stray radiations which would seriously impair the health of operators.

coolingwater VACUUM

cathodetungstentarget

X-RAY TUBEshield

copperanode

filament

glass tubex-rays

castingcavity

photographic film

result ingnegative image of cavity

Like light, X-rays travel in straight lines, but whilst metals are opaqueto light they are moderately transparent to X-rays, particularly those ofhigh frequency. Fig. 2.15 illustrates the principle of radiography. A castingis interposed between a shielded source of X-rays and a photographic film.Some of the radiation will be absorbed by the metal so that the density ofthe photographic image will vary with the thickness of the metal throughwhich the rays have passed.

2.92 X-rays are absorbed logarithmically—

/ = Ioe-*

Where I0 and / are the intensities of incident and emergent radiationrespectively, d the thickness and pi the linear coefficient of absorption ofradiation, (i is lower for radiation of higher frequency.

A cavity in the casting will result in those X-rays which pass through thecavity being less effectively absorbed than those rays which travel throughthe full thickness of metal. Consequently the cavity will show as a darkpatch on the resultant photographic negative in the same way that a greaterintensity of light affects an ordinary photographic negative.

A fluorescent screen may be substituted for the photographic film sothat the resultant radiograph may be viewed instantaneously. This type offluoroscopy is obviously much cheaper and quicker but is less sensitivethan photography and its use is generally limited to the less-dense metalsand alloys.

2.93 y-ray Methods can also be used in the radiography of metals. Sincethey are of shorter wavelength than are X-rays, they are able to penetratemore effectively a greater thickness of metal. Hence they are particularlyuseful in the radiography of steel, which absorbs radiation more readilythan do light alloys.

2.94 y-rays constitute a major proportion of the dangerous radiationemanating from 'nuclear waste' and from the fall-out of nuclear explosions.Initially naturally-occurring radium and radon (18.70) were used as asource of y-rays but artificially activated isotopes of other elements arenow generally used. These activated isotopes are prepared by bombardingthe element with neutrons in an atomic pile. A nucleus struck by a neutronabsorbs it and then contains an excess of energy which is subsequentlyreleased as y-rays. Commonly used isotopes are shown in Table 2.4. Ofthese iridium-192 and cobalt-60 are the most widely used in industry.

2.95 Manipulation of the isotope as a source of y-radiation in metallur-gical radiography is in many respects more simple than is the case withX-rays, though security arrangements are extremely important in view ofthe facts that y-radiation is 'harder' than X-radiation and that it takes placecontinuously from the source without any outside stimulation. All y-raysources are controlled remotely, generally using a manual wind-out system(Fig. 2.16). When not in use the isotope is stored in a shielded containerof some y-ray absorbent material such as lead. Because they are 'harder'than X-rays, y-rays can be used to radiograph considerable thicknesses ofsteel. Since the radiation source is small and compact and needs no external

Fig. 2.16 y-radiography manual remote wind-out system,(i) y-ray source exposed; (ii) y-ray source stored.

power supply, y-radiation equipment is very portable and can be used toexamine materials in situ, eg welded joints in motor-way bridges.

All forms of ionising radiation such as X-rays and y-rays are very harmfulto all living tissue and their use in the UK is governed by the FactoriesAct—The Ionisating Radiations (sealed sources) Regulations 1969'.

2.96 Ultra-sonic Methods In marine navigation the old method of'Swinging the lead' was used to determine the depth of the water underthe boat. This was replaced in the technological age by a 'sonic' methodin which a signal was transmitted from the boat down through the water.The time interval which elapsed between transmission and reception of the'echo' was a measure of the depth of the ocean bed. The ultrasonic testingof metals is somewhat similar in principle. Ordinary sound waves (of fre-quencies between 30 and 16 000 Hz) tend to bypass the small defectswe are dealing with in metallic components and ultra-sonic frequencies(between 0.5 and 15 MHz) are used for metals inspection.

When an ultrasonic vibration is transmitted from one medium to anothersome reflection occurs at the interface. Any discontinuity in a structurewill therefore provide a reflecting surface for ultrasonic impulses (Fig.2.17). A probe containing an electrically-excited quartz or barium titanate

(i)

(ii)

manualwind-out

flexiblecable-drive

isotope'safe'

probeV-raysource

radiationshielding

Table 2.4 y-ray sources used in industry

Isotope

Cobalt-60

Caesium-137

Thulium-170

Ytterbium-169

lridium-192

Symbol

60Co27

137Cs55

1 7 0 T-S 9

169Yb70

192Ir77

Half-life

5.3 years

33 years

127 days

31 days

74 days

Relative energyoutput in terms ofy-radiation

1.3

0.35

0.0045

0.021

0.48

Typical uses

50-200mm of steel

25-100mm of steel

2-13mm of aluminium

2-13mm of steel

10-90mm of steel

Fig. 2.17 The detection of a fault in plate material by ultrasonics.In (i) the impulse is reflected from the lower surface of the plate; whilst in (ii) it is reflectedfrom a defect. Measurement of the time interval between transmission of the impulse andreflection of the echo determines the depth of the fault.

crystal which can both transmit and receive high-frequency vibrations isused to traverse the surface of the material to be examined (Fig. 2.18.).The probe is coupled to a pulse generator and to a signal amplifier whichtransfers the resultant 'image' to a CRT (cathode-ray tube).

2.97 In satisfactory material the pulse will pass from the probe unim-peded through the metal and be reflected from the lower inside surface atA back to the probe, then acting as receiver. Both transmitted pulse andecho are recorded on the CRT and the distance, ti, between peaks is pro-portional to the thickness, t, of the test piece. If any discontinuity isencountered such as a blow-hole, B, then the pulse is interrupted andreflected as indicated. Since the echo returns to the receiver in a shortertime an intermediate peak appears on the CRT trace. Its position relativeto the other peaks gives an indication of the depth of the fault beneath thesurface.

Fig. 2.18 shows separate crystals being used for transmitter and receiverbut, as mentioned above, in many modern testing devices a single crystalfulfils both functions. Different types of probe are available for materialsof different thickness and this method is particularly useful for examiningmaterial—such as rolled plate—of uniform thickness.

Fig. 2.18 Basic principles of ultrasonic testing.The values of 'd' and T in the test piece are proportional to the values of 'di1 and V shownon the CRT.

metalplate

defect(ii)(i)

PULSEGENERATOR

SIGNALAMPLIFIER

TIME BASE

CRT

PROBE

TESTPIECE

transmittedpulse

echofrom B

echofrom A

Exercises1. Differentiate between:

(i) Malleability and ductility;(ii) Toughness and hardness;

(iii) Yield strength and tensile strength. (2.20)2. An alloy steel rod of diameter 15 mm is subjected to a tensile force of 150 kN.

What is the tensile stress acting in the rod? (2.30)3. Fig. 2.19 represents the force-extension diagram for:

(i) annealed copper;(ii) hard-drawn copper;

(iii) annealed low-carbon steel; or(iv) cast iron? (2.32)

4. When a steel wire 2.5 m long and of cross-sectional area 15 mm2 was subjectedto a tensile force of 4.0 kN, it stretched elastically by 3.2 mm. CalculateYoung's Modulus of Elasticity for the wire. (2.30)

5. A low-nickel steel in the heat-treated condition had an 'engineering' tensilestrength of 708 N/mm2. The reduction in area of cross-section at the fracturewas 44%. What was the true tensile strength of the steel? (2.31)

6. During a tensile test on a cold-worked brass the following figures were obtainedfor force and corresponding extension:

Extension (mm) 0.1 0.2 0.3 0.4 0.5 0.6 0.8 1.0

Force (kN) 23 46 69 82 89 94 102 110

Ext. (cont.) 1.5 2.0 2.5 3.0 4.0 4.3

Force (cont) 123 131 136 139 132 118 (Break)

The diameter of the test piece was 16 mm and the gauge length used was80 mm. Draw the force-extension diagram on squared paper and determine:

(i) Young's modulus of elasticity;(ii) the 0.1% proof stress;(iii) the tensile strength;(iv) The percentage elongation of the material. (2.30 and 2.32)

7. An aluminium alloy has a modulus of elasticity of 69 kN/mm2 and a yieldstrength of 275 N/mm2.

(a) What is the maximum force which a wire 3 mm in diameter could supportwithout suffering permanent deformation?

(b) If a wire of this diameter and 25 m long is stressed by a force of 430 Nwhat will be the elongation of the wire? (2.30)

forc

e

extension Fig. 2.19.

8. What method of hardness determination would be suitable for each of thefollowing components:

(i) a small iron casting;(ii) a large steel roll in situ;

(iii) small mass-produced finished components;(iv) a small hardened steel die.

Justify your choice of method in each case. (2.40)9. What important information is obtained from impact tests? (2.50)

10. What inspection techniques would be appropriate for detecting the followingdefects in cast products:

(i) internal cavities in a large steel casting;(ii) surface cracks in grey iron castings;(iii) surface cracks in aluminium alloy castings;(iv) internal cavities in aluminium alloy casting?

Give reasons for your choice of method in each case. (2.80-2.90)11. What non-destructive testing methods would be applied to reveal the presence

of:(i) subcutaneous slag inclusions in a thick steel plate;

(ii) quench-cracks in a heat-treated carbon steel axle;(iii) surface cracks near to a welded joint in mild-steel plate?

Give reasons for your choice of method in each case and outline the principlesof the method involved. (2.80-2.90)

BibliographyBateson, R. G. and Hyde, J. H., Mechanical Testing, Chapman & Hall.O'Neill, H., Hardness of Metals and Its Measurement, Chapman & Hall, 1967.BS 18: 1987 Methods for Tensile Testing of Metals (including aerospace materials).BS 240: 1986 Methods for Brinell Hardness Test.BS 427: 1981 Methods for Vickers Hardness Test.BS 891: 1989 Methods for Rockwell Hardness Test.BS 4175: 1989 Methods for Rockwell Superficial Hardness Test (N and T Scales).BS 131: 1989 Methods for Notched-bar Tests (Part 1-Izod; Part 2-Charpy).BS 3855: 1983 Method for Modified Erichsen Cupping Test for Sheet and Metal

Strip.BS 1639: 1983 Methods for Bend Testing of Metals.BS 3889: 1983 and 1987 Methods for Non-destructive Testing of Pipes and Tubes.BS 5996: 1980 Methods for Ultrasonic Testing and Specifying Quality Grades of

Ferritic Steel Plate.BS 4080: 1966 Methods for Non-destructive Testing of Steel Castings.BS 4124: 1987 Methods for Non-destructive Testing of Steel Forgings.BS 3923: 1972 and 1986 Methods for Ultrasonic Examination of Welds.BS 6443: 1984 Method for Penetrant Flaw Detection.BS 2600: 1973 and 1983 Methods for Radiographic Examination of Fusion Welded

Butt Joints in Steel.BS 3451: 1983 Methods for Testing Fusion Welds in Aluminium and Aluminium

Alloys.BS 709: 1981 Methods for Testing Fusion Welded Joints and Weld Metal in Steel.BS 6072: 1986 Methods for Magnetic Particle Flaw DetectionBS 4331: 1987 and 1989 Methods for Assessing the Performance Characteristics of

Ultrasonic Flaw Detection Equipment.