Powder Technology, 76 (1993) 23-30 23

Attrition of brittle spheres by fracture under compression and

impact loading

P.H. Shipway and I.M. Hutchings

Universiry of Cambridge, Depanment of Materials Science and Metallurgy, Pembroke Street, Cambridge CB2 3QZ (UK)

(Received January 11, 1993; in revised form April 1, 1993)

Abstract

Results are presented of a theoretical and experimental study of the fracture of single brittle spheres by uniaxial
compression between opposed platens and by free impact against plane targets. The stress distributions in elastic
spheres are broadly similar under both types of loading, with significant tensile components inside the sphere
on the axis of the system and on the surface of the sphere, around the equator for the case of compression.
The magnitudes and locations of the peak values of these stresses depend on the size of the contact area (relative
to the size of the sphere), which in turn depends on the mechanical properties of the target and platen material.
A simple equation used by previous investigators to estimate the maximum internal tensile stress from the load
provides a useful approximation only under certain conditions, and more generally leads to significant error.
Experiments with lead glass spheres have shown that under many conditions of compression and impact testing,
failure initiates on the surface, rather than internally, at a critical value of tensile stress. The work is relevant
to the breakage attrition of brittle particles in powder transport, handling and processing, and also to comminution.
It also provides a basis for the interpretation of indirect tensile tests in which spherical samples are uniaxially
compressed.

Introduction easily in a fixed testing environment. However, since

many different types of test apparatus can be used, it
Attrition, the unwanted breakdown of particles by is difficult to correlate results obtained in different
fragmentation or abrasion, occurs in many processes systems. Single particle tests, in contrast, can allow
where particles are used or transported. This results well-characterized loading to be applied to single par-
in the production of dust and may lead to loss of ticles, and their attrition behaviour under those con-
particles from the system; both these phenomena may ditions can be assessed. The role of abrasion in such
require costly remedial action such as shut-down of tests can be minimised, and they can therefore be useful
plant to replace the particles, or filtration to remove for determining the role of particle breakage in attrition.
the dust hazard. The inevitable broadening of the Single particle tests can be broadly classified as either
particle size distribution due to these effects may also impact or compression methods. Compression testing
cause problems in subsequent use of the particles. has been used by many workers to measure the load-
The attrition process is affected by many variables, bearing capability of particles. Sikong et al. [2] employed
involving both the particle and the system. It is an this type of test to assess the breakage behaviour of
extremely complex process, and the problem of mo-
fine particles of brittle minerals and coal, and used
delling attrition in a system has therefore usually been
eqn. (l), as given by Hiramatsu and Oka [3], to calculate
tackled by statistical methods, involving breakage func-
the stress, a, required to initiate fracture in a spherical
tions or kinetic formulations. A review of the literature
on the modelling of attrition and on suitable test methods particle of radius R from the crushing load, F,,.
has been given by Bemrose and Bridgwater [l]. Ex-
periments designed to study attrition can be divided uf=0.7F,,hR2 (1)
broadly into those tests involving single particles and
those which employ many particles. The many-particle Other workers have estimated the elastic strain energy
tests have the advantage of studying a system of particles required to cause fracture in a uniaxial crushing test;
which is statistically representative; the attrition tend- however, this method will only be valid if the fracture
encies of different particles may therefore be compared of the particles is governed by crack propagation rather

0032-5910/93/$6.00 0 1993 - Elsevier Sequoia. All rights reserved

24

than initiation, since in the latter case the elastic strain as shown in Fig. l(b), may be derived by the super-
energy may far outweigh the energy required to cause imposition of such a solution with its reflection in a
fracture. Rather less attention has been paid to single- plane perpendicular to the axis of the system and
particle impact testing, even though this can provide passing through the centre of the sphere. Expressions
a more realistic simulation of attritive breakage by for the stress distribution in a sphere subjected to
impact than the quasi-static compression test. uniaxial compression have also been derived indepen-
In the present paper, recent theoretical and exper- dently by Hiramatsu and Oka [3]. An approximation
imental studies by Shipway and Hutchings [4] of the to their solution, for the case where the contact area
fracture of single brittle spheres by compression and between the sphere and platen is small, has been used
by impact will be discussed. The results have direct extensively by subsequent investigators to calculate the
relevance to attrition processes involving the impact of
maximum internal tensile stress on the axis of compres-
brittle spheres, such as shot and bead-peening, and to
sion at failure for spheres under uniaxial loading,
applications in which spheres are subjected to uniaxial
irrespective of the actual size of the contact area. This
compression, (e.g. as ‘proppants’ used in petroleum
approximation suggests that the maximum tensile stress
extraction). It may also be relevant to particles in
packed beds, where they are subjected to multi-axial is N 0.7 times the applied load divided by the equatorial
compression, or to the flow of bed materials where cross-sectional area of the sphere.
impact with other particles, stirrers, bluff bodies or Kschinka et al. [6] tested glass spheres in compression
vessel walls provides the major source of breakage and inferred from an analysis of the failure stresses
attrition. by Weibull statistics that these spheres failed from
volume (i.e. internal) flaws. They suggested that the
internal tensile stress on the axis (a,) was therefore
responsible for fracture. Much greater tensile stresses,
Previous studies of the fracture of spherical particles a,, were present at the surface of the sphere just outside
by impact or compression the contact area but did not produce Hertzian-type
cone cracks in the spheres, although cracks of this type
The stress distribution in an elastic sphere subjected have been seen in other investigations of the compression
to compression or free impact against a plane target failure of brittle spheres [7-91; they appear to have no
has been studied in two important previous pieces of effect on the bulk failure of the specimen, but cause
work. Dean et al. [5] examined the case of an elastic only local chipping around the contact area.
sphere decelerated by a uniform pressure applied to Other workers have suggested that the fracture of
a single spherical cap. This geometry is illustrated in
brittle spheres in uniaxial compression occurs by a
Fig. l(a). Their solution is directly applicable to the
different mechanism. Tanaka et al. [lo] proposed that
case of free impact attrition.
in sintered ferrite spheres compression failure was
Since the elastic solution is linear, the stress distri-
bution in a sphere compressed between flat platens, initiated within the contact zone, possibly due to some
plastic flow within the particle. Kapur and Fuerstenau
[ll] and Arbiter et al. [7] both suggested that shear
deformation on conical planes leads to the opening of
cracks in the central region, followed by splitting of
the sphere. Failure by this mechanism might be expected
to initiate at a critical value of the maximum shear
stress within the sphere.
Arbiter et al. [7] conducted impact tests with massive
(74-124 mm diameter) brittle sand-cement spheres and
observed that fracture of the spheres, into wedge-shaped
segments, originated around a cone of densely com-
pressed material, the base of which was formed by the
contact area between the sphere and the target. How-
ever, Chaudhri [12], in a study of the impact fracture
of 1 mm spheres of glass and sapphire, suggested in
(b) contrast that the projectiles started to split along axial
Fig. 1. Geometry of loading in (a) free impact testing and (b) planes from the opposite end of the axis to the contact
uniaxial compression testing of spheres of radius R. area.
 25

Stress distributions in elastic spheres under impact

and compression loading

The solutions for the internal stresses in an elastic

sphere from the papers of Dean et al. [5] and Hiramatsu
and Oka [3] are given in corrected form by Shipway
and Hutchings [4]. In both cases, the stresses are
expressed as summations of infinite series; for numerical
computation, these must be truncated, and this was
done so that the computed value had converged to
within 1% of the true value. The solutions of Hiramatsu
and Oka [3] and of Dean et al. [5], extended to uniaxial 0 0.2 0.4 0.6 0.8 1
compression, lead to very similar numerical results. aclA
(a)
However, the solution of Dean et al. [5] offers much
greater computational simplicity than the more widely
used solution of Hiramatsu and Oka [3], and only the ~ Surface tensile stress
----_ Axial tensile stress
former solution will therefore be considered further.
In these solutions, it is assumed that the contact
pressure is constant over the contact area. For a sphere
which remains elastic and a platen or target which
undergoes plastic deformation during loading, this is
a good assumption. It is still valid if the sphere ex-
periences plastic deformation, provided that the extent
of plastic flow is small. If the sphere and the platen
or target both remain elastic throughout the loading
cycle, then the assumption will be lessvalid; nevertheless, 0 0.2 0.4 0.6 0.8 1
a,lR
the computed results will probably still provide a val- @I

uable insight into the distribution of the most important 20

stresses within the sphere.
Shipway and Hutchings [4] gave numerical values for
16
the tensile and shear stresses on the axis and the tensile
iLo
stresses at the surface, a+, for spheres under both l-d

compression and impact loading. Stress distributions “atz

were plotted along the axis of the sphere and along
2
a surface meridional line for various values of contact
radius, a,. The maximum values of the various stresses 2
were also given for a range of values of the contact
radius. To calculate the stresses at fracture, the load
F0 and contact radius a, must be known. In cases where
fracture of the spheres is governed by initiation rather 0 0.2 0.4 0.6 1
aclR o’6
than propagation of cracks, the peak values of the (c)

relevant stresses are critical. Figures 2(a)-2(c) show Fig. 2. Maximum values of stress within spheres: (a) due to
the maximum values of the stresses of interest for a impact: peak values of tensile stress on the axis (perpendicular
to the axis) and at the surface; (b) due to uniaxial compression:
range of values of a,/R for spheres of radius R under peak values of tensile stress on the axis (perpendicular to the
uniaxial compression and impact loading. The spherical axis) and at the surface; (c) peak values of shear stress on the
polar coordinate system used is defined in Fig. 1. The axis under impact and under compression loading.
maximum values of these three stresses for impact or
compression can be readily calculated from these graphs
and from the values of the force at fracture and the In impact testing, the initial velocity may be readily
contact radius at that point. measurable. For cases where a plastic indentation is
In compression testing, the force at fracture may be formed in the target, the radius of the spherical indent,
measured with a load cell, and the contact radius at a,, can be measured, and its depth, h, calculated from
fracture either measured, if residual plastic impressions the following formula:
are formed in the platens, or calculated from Hertzian
elasticity theory if the contact remains fully elastic. h=R-dw (2)
26

The volume of the indent, V, can be calculated as Compression platens and impact targets of a wide
follows: range of materials, both metallic and ceramic, were
used. The materials and their relevant mechanical prop-
v= 7Th2(3R-h) erties are listed in Table l*. All platens were ground
(3)
3 and polished to a 1200 grit finish; the platens were of
If a constant indentation pressure acts between the various sizes, but all were large compared with the
particle and target and the initial kinetic energy of the deformation induced in them by the compression or
impact is completely converted to plastic work in the impact testing.
target to create the indentation, then Impact testing was conducted with a gas-gun of the
type described by Hutchings and Winter [14] in which
pV= 0.5mU2 (4) particles are loaded into a sabot which is accelerated
along a tube by compressed gas; the sabot is stopped
where p is the indentation pressure, m is the mass of
at the muzzle and the particles continue freely to impact
the impacting sphere and U is its initial velocity. Thus,
the target a short distance away. For each test 10
the indentation pressure can be calculated, and since
spheres were fired at the target at an accurately mea-
the contact radius is known, the maximum force of
sured velocity (& 1%). The spheres or fragments were
impact can be computed.
recovered, and the number of spheres unbroken after
For target materials where no plastic indentation
impact was noted. Any damage to the target was also
forms, and no permanent deformation of unfractured
examined, and any permanent impressions were mea-
spheres is detected, the impact can be assumed to be
sured. Broken spheres were examined by scanning
elastic. An equation, derived from Hertzian theory,
relating the maximum load F,, during impact to the electron microscopy (SEM) to elucidate the mechanisms
impact velocity U has been given by Knight et al. [13]:
\0.6,= \ -0.4
TABLE 1. Properties of platen and target materials
/c
Platen or target Vickers Young’s Poisson’s
material hardness modulus ratio
where p is the sphere density. k is given by (GPa) (GPa)

Cower (1) 0.89 130

Aluminium alloy (I) 1.57 72
Titanium alloy (I) 3.25 110
where E is Young’s modulus, Y is Poisson’s ratio, and Mild steel (I) 1.98 210
Brass (I) 1.22 105
the subscripts 1 and 2 denote the particle and the
target respectively. From the maximum load calculated Aluminium 1.02 70
in this way, the maximum contact radius can be cal- Copper (C) 1.07 130
culated from the equation: Aluminium alloy (C) 1.77 72
Titanium alloy (C) 3.89 110
Mild steel (C) 2.78 210
ac3= i kFoR Stainless steel 2.00 210
Brass (C) 1.39 10.5
0.25% C steel 2.47 210
where a, is the contact radius. These calculations assume
0.4% C steel 2.81 210
that the sphere and target remain elastic and that any Zirconia (PSZ) 14.0 170 0.30
cone cracks formed in the target have no effect on the Soda-lime glass 6.14 70 0.25
pressure distribution across the interface. Tool steel 12.1 210 0.30
Silicon (single crystal) 9.29 100 0.27
Glass-ceramic (silicate) 12.1 130 0.21
Alumina (sintered) 16.1 380 0.24
Experimental method Silicon carbide (sintered) 30.5 410 0.17
Boron carbide (sintered) 36.3 450 0.17

The spheres studied in this work were of lead glass Vickers hardness of metals measured at 196 N load. Vickers
with diameters of 700+50 pm, supplied by Jencons hardness of silicon measured at 1.88 N. All others measured at
Ltd., Leighton Buzzard, UK (type H102/1 No. 6). The 4.92 N. (I) used as targets in impact testing only; (C) used as
platens in compression testing only.
glass had major constituents of 58% Si02, 25% PbO,
9% K,O, 4% Na,O and 2% B,O, (all percentages by
‘In some cases, materials with nominally the same composition
weight). Young’s modulus was 62 GPa, Poisson’s ratio used as compression platens and impact targets had different
was 0.25, and the Vickers hardness measured by in- hardnesses, since they had different microstructures; these ma-
denting polished cross-sections was 5.46 GPa. terials are clearly identified in the table.
 27

of fracture. A range of velocities was used for each The fragments of the spheres broken by impact tended
target material, leading to a velocity below which <5 to be hemispherical or in the form of wedge-shaped
of the spheres were broken and a velocity above which segments; sometimes a cone of densified material was
>5 were broken. The actual fracture velocity for the evident. Figure 3 shows a roughly hemispherical frag-
spheres against that target was assumed to lie between ment. Spheres broken by impact with plastically de-
these two velocities. This method of determining a forming targets showed much smaller fragments, and
‘fracture velocity’ was employed to take into account little information could be gained about their fracture
the statistical nature of fracture in the spheres. mechanisms.
In the compression tests a single sphere was loaded
to fracture in uniaxial compression between pairs of Compression testing
parallel platens in a screw-driven Schenck Trebel testing Considerable variation was seen in the fracture loads
machine. The load at which fracture of the sphere for the spheres compressed between different platen
occurred was accurately determined and experiments materials. In soft platens, the spheres formed plastic
with each combination of platen and sphere material indentations as the load was increased; with the softest
were repeated 20 times to reduce the error in the platens, the spheres became completely embedded in
calculated mean fracture load. Fragments of the frac- the platens and did not fracture at all. In cases where
tured spheres were retrieved for later examination. The failure of the spheres did occur, the load was noted;
platens were examined after use to investigate any the mean fracture load was calculated in each case
damage. Any residual plastic impressions were measured from twenty tests and is shown, together with the
by optical microscopy, and damage such as cracking standard error in the mean, in Table 3. The values of
of the more brittle platen materials was also noted. the internal and surface stresses of interest were cal-
culated for each platen-sphere combination, as de-
scribed above and these are also tabulated.
Results As in the case of impact failure, the sphere fragments
took the shape of segments or hemispheres. SEM
Impact testing micrographs of fragments of spheres crushed between
The results of the impact tests are listed in Table silicon carbide platens are shown in Fig. 4. In Fig. 4(a),
2. There was a considerable variation in the fracture a cone of densified material is shown, with the flattened
velocity; with a soft target, fracture of the spheres contact area forming its base. Figure 4(b) shows a
occurred at a relatively high velocity whereas the harder, wedge-shaped fragment of a glass sphere, with river
stiffer targets caused fracture at lower impact velocities. lines emanating from the area to the left of the mi-
The limits for the fracture velocity, as defined above, crograph from which a conical fragment may have been
were used to determine the values of the contact force detached.
using the appropriate elastic or plastic analysis as
outlined above. This information, together with the
calculated (in the case of elastic contact) or measured Discussion
(in the case of plastic indentation) contact radius at
fracture was used to calculate the stresses of interest The experimental work reported above has shown
at fracture, and these are also listed in Table 2. that under conditions of both impact and compression

TABLE 2. Results of impact experiments with 700 pm diameter lead-glass spheres

Target material Fracture Contact radius Contact pressure (q-u+)/2 (int) o+ (int) o+ (surf)
velocity (pm) (GPa) (GPa) (MPa) (MPa)

Glass 175-200 140-147 (E) 2.19-2.34 281-314 247-285

PSZ 41-48 73-78 (E) 1.54-1.62 105-115 55-63
Sic 47-82 75-93 (E) 1.82-2.34 129-202 65-128

Copper 250-288 251 (P) 1.48-1.97 0.55-0.73 89-118 142-188

Aluminium alloy 216-240 222 (P) 1.9s2.38 0.56-0.69 124-153 161-199
Mild steel 201-218 207 (P) 2.27-2.67 0.87-1.02 144-170 175-205
Brass 216-289 235 (P) 1.50-2.68 0.57-1.01 95-170 134-240
Titanium alloy 174-203 185 (P) 2.75-3.74 1.04-1.42 165-224 182-248

(E) contact radius and contact pressure calculated from elastic theory (see text). (P) contact radius measured, and contact pressure
calculated from plastic theory (see text).
28

the inevitable presence of surface flaws. The fact that

the surface stress at fracture was in most cases only
slightly lower than the internal tensile stresses therefore
adds weight to the suggestion that in these cases failure
was initiated at a critical value of the surface stress,
of about 200 MPa.
Spheres which fractured against elastically deforming
targets, or in compression between elastic platens, in
contrast, showed a wider range of surface tensile stress
and a much more constant value of internal shear stress
(around 2 GPa). Although the shear strength of these
glass spheres is not known, some estimate can be made
Fig. 3. Fragment of glass sphere fractured by impact against from studies of other silicate glasses, and a value of
partially stabilized zirconia target. c. 2 GPa seems reasonable [15]. For the spheres frac-
tured against elastically deforming surfaces (except for
failure, the glass spheres tended to fracture into wedge- the impact tests against glass), the values of peak surface
shaped segments, although it was not possible from tensile stress at failure were always less than about
the morphology of these fragments alone to deduce 200 MPa, and therefore too low to cause fracture from
the mechanism of failure. Where the target or platen surface flaws. This suggests that in these cases fracture
suffered only elastic deformation, conical glass fragments was initiated by local internal shear failure rather than
(as seen in Fig. 4(a)) were also sometimes found. These by surface tensile failure, although propagation of the
latter features are consistent with failure of the spheres cracks once initiated would almost certainly have been
by internal shear. affected by the internal tensile stresses. In the case of
Examination of the values of the various stress com- the spheres fractured by impact against the glass target,
ponents at failure, listed in Tables 2 and 3, suggests both the internal shear stress and the surface tensile
that under impact or compression loading against plast- stresses appear to have been high enough to cause
ically deforming surfaces, the maximum values of the failure, and a unique failure mechanism cannot be
surface tensile stress show appreciably less variation deduced.
than those of the internal tensile stress. This alone In much of the earlier work on the quasi-static
suggests that the former is responsible for failure. It compression failure of brittle spheres, the experimental
is also relevant to note that for glass, the surface results have been apparently successfully analysed by
strength would be expected to be much lower (by a assuming that the sphere had been split by a tensile
factor of >20 [15]) than the internal strength, due to stress acting at its centre, and that this stress was -0.7

TABLE 3. Results of uniaxial compression tests on 700 pm diameter lead-glass spheres

Platen material Fracture load Contact radius (q-u@ (int) a+ (int) 04 (surf)
(N) (pm) (GPa) (MPa) (MPa)

B,C 160*7 90 (E) 2.28 248 172

Sic 167+7 92 (E) 2.30 257 178
Alumina 167*9 92 (E) 2.29 257 177
PSZ 160*9 95 (E) 2.05 244 171
Tool steel 173f7 96 (E) 2.18 262 183
Glass-ceramic” 168f8 100 (E) 1.97 253 179
Silicon” 171*10 103 (E) 1.89 256 183
Glass” 173*11 109 (E) 1.72 254 181

Titanium alloy 208*9 136*3 (P) 1.40 284 218

0.4% C steel 191 f 12 154*4 (P) 1.06 246 196
Mild steel 201 f 10 161*4 (P) 1.04 251 207
0.25% C steel 194+11 167*5 (P) 0.94 234 198
Stainless steel 197* 11 173*5 (P) 0.95 242 209
Aluminium alloy 227 + 12 205 rt 5 (P) 0.84 217 223
Brass 1961t 14 207 f 6 (P) 0.72 185 190

The spheres became completely embedded, and no fracture occurred, with the copper and aluminium platens.
‘Platens exhibited Hertzian cone cracking. (E) indicates that the contact radius was calculated from the elastic theory. (P) indicates
that contact was plastic and that the contact radius was measured.
 29

failure mechanisms would also be expected for other

brittle materials which, unlike glass, contain bulk flaws
which are comparable in size with the surface flaws.
Thus, the use of uniaxial compression to derive the
tensile strength of spherical samples is valid in some
cases and the doubts cast upon the method by Darvell
[16] can be allayed. However, it must be noted that
in some cases the method provides a measure of the
surface tensile strength, rather than the bulk strength,
as has been previously assumed, unless the material is
insensitive to surface flaws. Even then, care must be
taken in interpreting the results of such indirect tensile
tests, since shear rather than tensile failure may occur.
(a)

Conclusions

The stress distributions in elastic spheres subjected

to free impact or uniaxial compression loading are
broadly similar, with significant tensile components
inside the sphere on the axis of the system and on the
surface of the sphere, and with the maximum values
of these stresses occurring around the equator in the
case of compression. High shear stresses are also present
Fig. 4. Fragments ofglass spheres fractured in uniaxial compression on the axis within the sphere. The magnitudes and
between silicon carbide platens. locations of the peak values of these stresses depend
on the size of the contact area, which in turn depends
times the applied force divided by the equatorial area on the mechanical properties of the target or platen
of the sphere. This assumption derives from the work material.
of Hiramatsu and Oka [3]. Darvell [16], however, in The simple equation (see eqn. (1)) used by previous
a comprehensive review of indirect methods of meas- investigators to estimate the maximum internal tensile
uring tensile strength, criticised this assumption, and stress from the load is a useful approximation only
concluded that “it is extremely doubtful if diametral under certain conditions, and more generally leads to
compression gives a tensile strength”. significant error. Experiments with lead glass spheres
As Fig. 2(b) shows, the factor of 0.7 is an arbitrary have shown that under conditions of compression and
value, and for values of a,/Rmuch greater or less than impact testing against plastically deforming surfaces,
0.2, it provides a poor approximation. In contrast, failure initiates on the surface of the sphere, at a critical
however, the tensile stress on the surface of the sphere value of tensile stress. This stress is approximately
at the equator remains almost constant for a given 0.4FohR2. With elastically deforming platens, failure
applied force over a large range of contact areas. This is initiated by internal shear. For materials with sig-
observation, coupled with the likelihood that in many nificant bulk flaws, failure may be initiated by the
cases fracture will have initiated at the surface rather internal tensile stress.
than internally, provides an explanation for the apparent
success of the approximation used in earlier work. It
should be noted that a factor of -0.4, rather than 0.7
as used above, yields a more accurate value for the
true tensile strength of the material, and that the method
will only be applicable to cases in which the sphere Acknowledgements
does indeed fail from surface flaws and not by internal
shear; that is, it will only be valid where the load is This work was supported by BP Research, BP In-
distributed over a substantial area of the sphere by ternational Ltd., and the UK Science and Engineering
contact with a plastically deforming platen. Different Research Council.
30

List of symbols References

C.R. Bemrose and J. Bridgwater, Powder Technol., 49 (1987)

radius of the circular contact area between sphere 97.
L. Sikong, H. Hashimoto and S. Yashima, Powder Technol.,
and platen or target
61 (1990) 51.
El Young’s modulus of sphere material Y. Hiramatsu and Y. Oka, Znt. Z. Rock Mech. Min. Sci., 3
E2 Young’s modulus of target material (1966) 89.
P.H. Shipway and I.M. Hutchings, Philos. Mug., ,467 (1993)
FO normal load on sphere at fracture
1389.
h depth of final indentation into platen or target W.R. Dean, I.M. Sneddon and H.W. Parsons, in Selected
k reduced elastic constant for sphere and target, Government Research Reports, Volume 6, Strength and Testing
of Materials, Part ZZ,Testing Methoa!s and Test Results, HMSO,
given by eqn. (6)
London, 1952, p. 212.
m mass of sphere 6 B.A. Kschinka, S. Perrella, H. Nguyen and R.C. Bradt, .Z,
R radius of sphere Am. Ceram. Sot., 69 (1986) 467.
7 N. Arbiter, C.C. Harris and G.A. Stamboltzis, Trans. AZME
u impact velocity of sphere
244 (1969) 118.
V volume of indentation into platen or target 8 H. Rumpf and K. Schonert, DECHEMA Monogr., 69 (No.
1292.1326) (1972) 51.
9 R.J. Verrall. J. Dent., 4 (1976) 11.
Greek letters 10 K. Tanaka, H. Yamamoto, S. Yumoto, T. Ichinose, T. Iimura
and H. Harada, .Z. Mater. Sci Lett., 4 (1985) 184.
CL mean contact pressure on surface of sphere 11 P.C. Kapur and D.W. Fuerstenau, Z. Am. Ceram Sot., 50
Vl Poisson’s ratio of sphere material (1967) 14.
Poisson’s ratio of target material 12 M.M. Chaudhri, in C.R. Kurkjian (ed.), Strength of Inorganic
v2
Glass, Plenum Press, New York, 1985, p. 87.
P density of sphere material 13 C.G. Knight, M.V. Swain and M.M. Chaudhri, .Z.Mater. Sci.,
“.P principal stress in the C#Idirection 12 (1977) 1573.
tensile stress in sphere at fracture 14 I.M. Hutchings and R.E. Winter, Z. Phys. E, 8 (1975) 84.
Uf
15 A. Kelly and N.H. Macmillan, Strong Solids, Clarendon Press,
UC3 principal stress in the 8 direction Oxford, 1986.
fT principal stress in the r direction 16 B.W. Darvell, Z. Mater. Sci., 25 (1990) 757.