46.

Progress of Glide and Fracture in Alloy Crystals 141

The behaviour of magnesium crystals is shown in Fig. 108. The
mean ultimate glide strain decreases clearly with increasing alloy
content, a phenomenon due to the growing tendency to glide fracture
along the basal plane. In order to assess technologically the effect
of alloying, one has to bear in mind not only the increase in the
critical shear stress of the pure metal (VallJ but also the subsequent
capacity for work hardening. Accordingly the total hardening
which can be achieved by alloying and glide to the point of fracture
(V all.'
v·Clr -- So, Se, all.
S all . SS· = Se, all.) is plottcd in Fig . 109 as a func-
o 0, all. 0
tion of the concentration.
In both cases it will be found that it is not the maximum but a

Z J V of 5 8
Atomic-% Al 01' Zn.
FIG. 109.-Work-hardening Capacity of Mg Solid Solutions (231).
-- Al-Mg, ---- Zn-Mg.

medium concentration to which the maximum value of this total

hardening corresponds. The cause of this must no doubt be sought
in the increase of brittleness, already referred to, which accompanies
an increase in alloying content. The work of deformation behaves
in a similar way to that of the product Vall .. VTr'
The relationship between the yield-stress curves and the con-
centration of cubic face-centred zinc-aluminium solid solutions is
shown in Fig. 1l0. The hardening caused by zinc additions,
particularly noticeable at the start of extension, is apparent from
these curves. The lattice rotation which accompanies stretching
of the crystal corresponds to the normal case of octahedral glide
which was described in Section 36. At higher zinc contents the
symmetrical position of the crystal axis, which leads to double glide,
142 Plasticity and Strength of Metal Crystals
is not attained, because the crystals had fractured within the region
of simple glide without necking. This fracture appears to be a
sort of shear fracture on the operative glide plane, similar to that
observed with magnesium solid solutions.
The yield-stress curves of et-brass crystals of various zinc con-
centrations are shown in Fig. 111. In spite of the initially increased
yield stress, the shear stress in the operative octahedral system
remains during extension below that of the pure copper crystal.
The higher the zinc content the lower is this shear stress. The
lattice rotation which accompanies glide again corresponds essen-

/'1 9

12 8
~
1 ~
1 ~
7

6
""~ ""~ 5
.,
..,
.,
.., ffs 72

-;z•.
'f
".,
..!:; " J.
ti 'I

"'""
1J.l

0 '10 80 fRO 0 10 20 30 'f0 50 fiO 70 80 90

Glide strain % Glide strain %
FIG. llO.-Yield-Stress Curves of FIG. 111.- Yield-Stress Curves of
Zn-Al Solid Solutions (239). a-Brass Crystals (234).

tially to the typical case for octahedral glide. The differ-

ences relate only to the interaction between the two operative
glide systems. The higher the zinc content, the more will the
exclusive operation of the first system extend beyond the sym-
metry position (Fig. 112). This means that the more zinc there
is in the crystals the more will the hardening of the initially latent
octahedral system exceed that of the operative system. Fig. 113
gives an example for the variation of the critical shear stress in the
two glide systems. The second system becomes active when its
resolved shear stress exceeds that of the first system by 30 per cent.
In the course of further extension the symmetry position (equality
of both shear stresses) is again exceeded, this time in the opposite
direction because it is now the first system, temporarily latent, which
experiences the greater hardening. Only under the influence of a
 46. Progress of Glide and Fracture in Alloy Crystals 143
shear stress that is 10 per cent. higher than that of the operative
system will it again become active. The higher the shear stress at
the start of glide, the less marked will be the subsequent hardening
with increasing extension. The new bands which appear with
each change in the glide system are shown in Fig. 114. The ultimate
tensile stress of crystals with 72 per cent. copper is between 12·9
and 32·9 kg.jmm.2, the elongation between 67 and approx. 122 per
cent.-according to their initial orientation (233).

of. Glide plane

xt., "

o !i0 fOO f!iO ZOO

Extension %
FIG. 112.-Unequal Rate of Hardening of FIG. 113.-Variation of the
the Operative and Latent Octahedral System Critical Shear Stress in the
in the Extension of a-Brass Crystals, Double Glide of a-Brass
Expressed in Helation to Concentration. Crystals with 70% Cu (240);
cf. also (233).

Copper-aluminium solid solutions exhibit a behaviour substantially

analogous to that of the O(-brasscrystals (24]). In this case, too, if
the initial shear stress is high the yield-stress curve soon intersects
that of the pure copper crystal, and exhibits for the higher glide
strains lower shear stresses than this. Here, too, the lattice rotation
of the alloy crystals is characterized by a substantial transgression
of the symmetry position, thereby revealing the greater bardenjng
of latent glide systems.
The extension of silver-gold solid solutions differs only compara-
tively slightly from the bebaviour of the pure metals (235). Only
at small deformations does the hardening of the solid solution remain
noticeably below that of the pure metals. The lattice rotation, too,

•
144 Plasticity and Strength of Metal Crystals
is little affected by the formation of solid solutions. The hardening
of the latent octahedral system in this case exceeds that of the active
system by about 12 per cent. at the most (at medium concentrations).
The copper-nickel series of solid solutions appears to behave in a
similar manner, i.e., there is at medium concentrations more pro-
nounced hardening of the latent octahedral system (236).
The significance of the atomic arrangement for shear hardening in
plastic extension is evident from Fig. 115. Whereas the cubic

b
FIG. 114 (a) and (b).-Change of the Operative Glide
System when Extending a-Brass Crystals (234).
Appearance of New Bands.
(a) After the first change; (b) after the second change.

face-centred AuCua crystal with disordered atom arrangement

0
(quenched from 800 C.) shows the type of yield-stress curves
characteristic of solid solutions (high initial shear stress and rather
slow increase of the shear stress with increasing glide), in the case
of ordered atom arrangement (produced by an ageing treatment of
0
the crystal at 325 C.) the hardening curve rises steeply with increas-
ing glide strain from a low initial value-as with pure metals. The
extent, too, to which the symmetrical position is transgressed,
corresponds to this behaviour: with disordered atomic arrangement
it is twice that of the ordered arrangement.
The extension of aged copper-aluminium solid solutions (with
 47. Influence of Temperature and Speed upon Start of Glide 145
5 per cent. eu) can also be represented by a mean hardening curve
(238). However, the fracture of these crystals does not obey a
condition of maximum load, as
in the case of pure aluminium. 16
Instead, as with Zn-Al solid
solutions, fracture occurs in cer- 1¥
tain orientation ranges, without
previous necking, along one or
more nearly plane surfaces. The
mechanism of this fracture is not
known. The ultimate tensile stress,
according to the initial orienta-
tion, lies between 33·0 and 45·5
kg.jmm.2, the elongation between
7 and 26 per cent.

(c) INFLUENCE OF TEMPERATURE

AND TIME
--
The description of the effect of
composition and structure upon
the glide of metal crystals will
now be followed by a discussion of
tU If{} DU
the significance of temperature Glide stmin %
and speed of deformation. The FIG. llo.-Yield-Stress Curves of
dynamic properties have hitherto AuCu3 Crystals (237).
referred to room temperature and
to a limited range of deformation speeds. However, it has been
precisely the experiments conducted over a wide range of tempera-
ture and time which have led, if not to an explanation, then at
least to a physical interpretation of crystal glide.

47. The Influence of Temperature and Speed upon the Start of

Glide
The influence of the test temperature on glide has hitherto been
widely investigated mainly with hexagonal crystals. It can be
said of them in general that the sharp bend in the stress-strain
curve at the yield point becomes more accentuated with rising
temperature. It was seen in Section 43, in which we discussed the
dependence of the stress-strain curve upon the orientation of the
crystal, that the sharpness of this bend was decisively influenced
also by the position of the glide elements. Where the glide elements
are more or less transverse to the direction of tension, an initial
L

•
 146 Plasticity and Strength of Metal Crystals
reduction in load with increasing extension is possible in spite of
the shear hardening due to glide, whereas the same yield-stress cW've

.,
(:l

8
"ij
;:e
c
.~
.,'"., ., 80
'"
'"
•.'"
'"c c
i;
T~523°obs.
<; 150
.~ ~
.~ .~ (:l
.,.,
., .,
:s
9'
1l
"5 100 ;:e-
;:e c (:l
;:e
c
~~.~
~ .,'"
1 50
Z'"

'"
.~
'"
..,.
;:e
'"'"
.,•.
'"'" .. ~
tJ:i
i/.i 0 J
Extension % Extension %
a b

., .,
~ •.
(:l

~<:l
~~;:e
C
.,.:;::'Q.c.,.
Col. FIG.1l6 (a)-(c).-Initial Parts
of the Stress-S train Curves
.t~ of Cd Crystals at Various
~~
.,. C
Temperatures (255). Xo angle
.~t of inclination of the basis.
The stresses used for deter-
"'- mining the yield point are
~.~
iI5:§ indicated by a O.
'0 1 g .1 5
Extension %
c

leads at once to a steep increase in load if the initial position of the

glide elements is very oblique.
A characteristic example of the initial parts of strcss-strain curves
of several fractions of the same cadmium crystal, obtained at various
temperatures, is found in Fig. 116, a, band c. The test temperature
ranged from 20° to 5850 K., i.e., to within go of the melting point.
A stress-strain curve obtained with special equipment at still lower
 47. Influence of Temperature and Speed upon Start of Glide 147
temperatures down to 1·2° K. is shown in Fig. 117. It is seen that
considerable plasticity persists even at these temperatures and that
in particular the critical shear stress is ofthe same order of magnitude
as at the other temperatures. This fact has also been confirmed for
zinc crystals [(251), (253)].
Fig. 118 shows in diagrammatic form the relationship between
critical shear stress and temperature. It will be observed that the
shear stress of the basal glide system of hexagonal metals is only
slightly affected by temperature. The increase in the shear stress

,-
v/
....,,'"
./
'1,20."<
.,,r

//
~
,/
5 '100
1,20
i

!
'"
.~
ZOO

~
,/

I
I
I i
\ i
70 5fl 60 70

Extension %
FIG. 117.-Stress-Strain Curve of a Cd Crystal at Low Temperatures (254).

of cadmium, for instance, is slightly less than four-fold, despite a

reduction in temperature from just below the melting point to
20° K. A similarly slight dependence of So on temperature has been
observed with the rhombohedral bismuth (263) and tetragonal tin
(262).
Although the experimental data available for very low tempera-
tures are scanty (a serious deficiency, especially where cubic crystals
are concerned), extrapolation down to absolute zero for zinc and
cadmium can at least be attempted. It reveals that even at this
temperature the low yield stress of the basal plane persists.
In the neighbourhood of the melting point there was in all cases
a zone of approximately constant critical shear stress. A decrease
148 Plasticity and Strength of Metal Crystals
to zero value (that of the melt) must proceed very suddenly, at any
rate for cadmium and bismuth; down to temperatures of go or 6°
below melting point there is still the same value as in the range in
which there is no dependence on temperature. In view of the
temperature dependence of the solubility of impurities, their role
in the above tests cannot at present be assessed.
It has already been mentioned that aluminium crystals show no
sharp yield point at room temperature (see Section 40). This has
also been observed for the temperature range -185 to 600° C. (257).
However, the yield-stress curves shown in Fig. 128 indicate that

2q()
\

~ 200
\
-f.. \
\
,;
"-
...
;:: 180
.,
',.
:;""0 "-
.8i

..
\.
~., f20
Zn
", --(g \..
~
B
., 80
.......•....
~ ,', __ u,

'"
t:>
-''£>

]
X

------ --E!... --;;--

----0--
00

~ '10
------ I-- )(";""-r"

+_x
o 100 200 300 '100 500 EiOO~
Absolute temperatw'e
FIG. llS.-Effect of Temperature on the Critical Shear Stress.

for aluminium, too, there is only slight dependence of the start of

glide on temperature. Whereas at 600° C.a shear stress of 100 g.jmm.2
suffices for perceptible deformation, at -185° C. a stress of approxi-
mately 800 g.jmm.2 is necessary to produce observable octahedral
glide.
Having described the effect of temperature, we will now deal
briefly with the effect of the speed of deformation on the critical
shear stress. It will be explained later (Section 50) that, owing to
their similar influence, time and temperature are probably very
closely connected. It has been found that in general the yield point
increases with increasing speed of testing. It is probable that the
decisive factor is not the increase in the stressing speed within the
range of purely elastic deformation, but the greater speed of deforma-
tion within the narrow range that lies between the elastic limit and
 48. Progress and Termination of Glide 149
the yield point. Some idea of the magnitude of the influence of
this speed can be gained from Fig. 119 in the case of cadmium
crystals. A hundred-fold increase in the speed of testing leads,
within a wide range of temperature, to an increase of the critical
shear stress of 20-30 per cent. Fig. 119 reproduces, also, as a
broken line, the curve which appears in Fig. 118; since it relates to
crystals which were drawn from the melt much more slowly, it is
not directly comparable with the other two curves.
It is difficult to give figures for aluminium crystals, owing to the
absence of a sharp yield point in this case. From available tests

100

80
'\ ,
f
<:i, 50
.~
'"'"
'" '1-0
~
1;
~'"
Qj
ZO

Sm
0 100 ZOO .lOO '1-00 soooK 500
Absolute tempe"attt1'e
FIG. 119.-Dependence of the Critical Shear Stress of Cd Crystals upon
the Speed of Stressing (Deformation) (252).

in which the test speed varied from 1 : 23,000 it appears that, in

spite of this wide range, glide begins at approximately the same
stress (256).

48. Progress and Termination of Glide in Relation to Temperature

and Speed
An example of the energy needed to continue extension at various
temperatures is shown in Fig. 120. It represents the extension of a
cadmium crystal in stages at various temperatures. The change in
temperature was always effected after removal of the load from the
crystal. It will be noticed that the relatively slight dependence of
the critical shear stress upon temperature in the initial unworked state
persists at larger glide strains. On the other hand, the work needed

•
150 Plasticity and Strength of Metal Crystals
to continue the extension is strongly dependent upon temperature.
Whereas at room temperature and even more so at elevated tempera-

:;l JZIJ
!:l
1y
] Z3IJ
.~
zIl.01
.
'"
~ Zfll
/

~ / x

l: I
'" ZIlIJ
I
~ A ~
.~
.:;:;'"
tolJ
37\."" I
I
I
S 1ZIJ
~
1 III ~- Z91° If ij
~ ,1'-<_
~~
Z91°
. "-X-x
!""--'" ~ a
,,;,
! I u
;
.~ 'Ill

~
"::l
!J
i t
r10 ZIJ
i
1
Z'I ZI JZ J6

Extension %
FIG. 120.-Stress-Strain Curves of a Cd Crystal at Various
(Absolute) Temperatures (254).

tures stretching can take place initially with a falling load (however,
here too the shear stress in the operative basal glide system
increases) a very considerable increase in load is necessary at low

FIG. 121.-Local Glide of a Cd Crystal Extended at 2500 C. (255).

temperatures. Consequently at low temperatures stretching usually

proceeds much more uniformly along the whole crystal than at
temperatures at which, owing to the reduction in stress, it is confined
initially to the weakest portions of the crystal which experience the
first extension (Fig. 121).
 48. Progress and Termination of Glide 151
As a rule the glide bands become coarser as the temperature rises
(Fig. 122 compared with Fig. 35a). Quantitative statements
regarding the mean thickness of the glide packets are not available.
Difficulty arises frol11the circumstance that, with increasing magnifi-
cation, new glide lines becomes visible. In the case of regular zinc

FIG. 122.-Zn Crystal Extended at 3000 C. (252).

polyhedrons which were grown from vapour, values were obtained

which fluctuated around 0·8 i.L., or a multiple thereof (267). Tin
crystals usually show glide bands; sometimes, however, they appear
perfectly smooth in spite of substantial deformation (extension up
to ten times the initial length); the thickness of the glide packets
must then be below microscopic resolution (261). Reduction of the

FIG. 123.-Formation of Coarse Glide Packets and of a

Shear Fracture in the Course of Very Slow Extension
of a Sn .Crystal (see 260).

speed of straining, like an increase in temperature, produces coarse

glide bands. Fig. 123 contains two illustrations of zinc crystals
which have been stretched extremely slowly.
The yield-stress curves obtained for hexagonal metal crystals at
different temperatures are shown in Fig. 124. They reveal a very
pronounced influence of temperature on the shear hardening of the
basal plane. The hardening coefficient derived from the slope of the
yield-stress curve changes 400-fold within the temperature range
152 Plasticity and Strength of Metal Crystals
investigated. This dependence is most marked at intermediate
temperatures; it decreases rapidly at both high and low tempera-

2800 r
2'100

~
•. 2000 ~ ~II
I

~
.t
.;, 1500
I
.~., J I ~
'.," 1200
oS
•..
I 1°:1
\~
<:j
I 2~Oo,
~'" 800
/,,1

-------
"-.l //
20°

'100
I~
n
JV
-
/'
'50°
00°

o /lOO 1100 000 800 fOOO f/lOQ

Glide strain %
(a) Magnesium (249).

1'1PP

..
~ tIltJtJ'

g-
';'/JOP
.,
.~
~ 5117
~
1?(J°
~ /'/pp
S}
2500

wo 800 12tJO 1000 .3tJtJtJ

Glide strain %
(b) Zinc (253).

tures, and becomes insignificant where these are extreme. For very
low temperatures this is apparent from Figs. 117 and 120 (cadmium
crystals).
 48. Progress and Termination of Glide 153
The amount of basal glide that can occur at different temperatures,
which in the case of magnesium is limited by shear fracture along the
basal plane, or by the start of pyramidal glide, and in the case of
zinc and cadmium by secondary glide in twin lamella) (see Section

7800

7'100
-253°
I I
FIG. 124 (a)-(c).-Yield-Stress Curves
1200
of Hexagonal Crystals as a Function

I I:
I
of Temperature.

80-100 g.fmm. /min.

2
Speed of Stressing
before Reaching the Yield Point:
Discontinuous
•• 1000 Curves Obtained at a Stressing Speed
I
1
t5,
, I
which was about 100 Times Greater.

.~ 800 r
~
~ I
6 600
I
~

'1-00

200

2000

o cOO If.00 600 QOO 1000' 1200 7'100

Glide strain %
(c) Cadmium (252).

44), is again well expressed by a limiting shear stress which is in-

dependent of orientation. Thus within a wide range of temperature
the mean yield-stress curves for basal glide are characterized not
only by their starting point and slope, but also by the point at which
they end. The relationship between temperature and the limiting
values for shear stress and glide strain is shown in Fig. 125. In the
temperature range investigated extension by basal glide increases

•
 154 Plasticity and Strength of Metal Crystals
by an order of magnitude, while at the same time hardening decreases
equally rapidly. This entails even with the flattest yield-stress

2800

2'100
2000

~ 2000
,';; "•..
.""e
03
1500
'"•..
MOO
e ~'"
'"'"'"'" t200
~ t200

."""" .~
""
.~
~
" 800
.~ 800
~
."" .""
>-::i >-::i
'100 '100
0_"4
" .•.

I
o tOO,p/OO SOO '100 500 fiOO· 0 100 200 SOO 'l{J0 500 6'(JIJ
Absolute temperature Absolute temperature
(a) Limiting shear stress. (b) Limiting glide strain.
FIG. 125 (a) and (b).-Termination of the Basal Glide of Hexagonal Crystals
as a Function of Temperature.

curve an increase in the shear stress during extension of three times

the original value. I
While the slope of the yield-stress curve and of the limiting glide

.-
5,0
"
."" "'fig
.~"
c
3,0
.

~""
~:?
'" e ~2
~o ,1Zn
'& 45
"< R- Cd
i:3
:= 0 tOO cOO SOO 'lOO 500 600'

Absolute temperatu1'e
FIG. 126.-Work of Deformation Required for Basal Glide as
a Function of Temperature.
(For Cd: X = slow fracture, in about 20 min.; 0 = rapid
fracture in about 10 sec.)

strain and shear stress markedly depend upon temperature, the

work of shear for the three hexagonal metals for basal glide is found
to be approximately independent of temperature (Fig. 126). Zinc
crystals, however, show a fiat maximum at 1500 C. It is not known
 48. Progress and Termination of Glide 155
for certain whether this is bound up with a possible deviation from
pure basal glide which may occur in this range of temperature.
The magnitude of the mean shear energy amounts to 4·4 cal./g. for
magnesium (that is 18 times the specific heat, or 1/10th of the latent
heat of fusion); for zinc it is 0·97 cal./g. (11 times the specific heat,
or 1/24th ofthe heat of fusion), and 0·24 cal./g. for cadmium (4 times
the specific heat, or 1/44th of the latent heat of fusion). This
purely empirical finding can hardly represent a limiting condition,
because in the course of deformation the greater part of the added
energy is transformed into heat and is therefore no longer present
in the crystal at the conclusion of glide. It would be more reason-
able to suppose a saturation with internal energy, which in the cases
described here, of course, would have to amount always to the same
fraction of the total work of deformation. (For further details see
Section 60.)
In Fig. 124, a and c, the dependence of the yield-stress curve on
time at different temperatures is also shown by curves obtained at a
one hundred-fold speed of deformation. It is found that within
the range in which the yield-stress curve is very dependent upon
temperature there is also a marked dependence upon speed; in the
case of magnesium, for instance, the shear stress after 600 per cent.
0
glide at 250 C., in the crystal that has been stretched rapidly, is
three times greater than in that which has been stretched slowly.1
Within the range in which the yield-stress curve is less dependent
upon temperature the speed of deformation is also without con·
siderable influence. The work of shear, which is independent of
temperature, is also largely independent of the speed of deformation,
as is seen in the case of cadmium in Fig. 126.
In the extension of zinc and cadmium crystals represented in Fig.
124, band c, the primary basal glide is restricted by the start of a
new basal glide within the twin lamellro formed towards the end of
primary glide. If the temperature exceeds the maximum values
given in the diagram, then the normal course of basal glide is dis-
turbed by re crystallization during the extension. At temperatures
above 300 C., cadmium crystals break transversely without con-
0

siderable elongation. With zinc crystals, decrease of temperature

leads at -80 C. to pronounced reduction of the ductility and to the
0

appearance of cleavage fracture along the basal plane-a phenomenon

1 Great importance therefore attaches to the constancy of test speed in
certain temperature ranges if the tests are to be comparable. In accordance
with the crystallographic process of glide, the value to be maintained constant
is the rate of glide strain which is independent of the initial position and of
the change in orientation in the course of stretching (256).

•
156 Plasticity and Strength of Metal Crystals
which is observed at all lower temperatures. (For further details
see Section 53.)
Only approximate details are available for tin crystals. While the
slope of the yield-stress curve remains more or less constant, the
limiting glide strain falls appreciably with rising temperature (up
to 200° C.). In this case the range within which the hardening

tP
",..-- -1!S'.

/
/
1/
/
7
----
f!'

~
I
'f
/
V
-- 1PP'

ePP'
~!
~
"'"~
."" V
/-
lPO'

~
,---
/" I
I
IPO'
,00'
6'PO'

6'0 Jp lOP 1tO 110 /6'P /N

Extension %
FIG. 127.-Stress-Strain Curves of Al Crystals in Relation to
Temperature (257).

coefficient is only slightly dependent on temperature is reached

already between 20° and 200° C. (262).
The relationship between temperatures and the stress-strain
curves of aluminium crystals is given in Fig. 127. From these it
will be seen that hardening is greatly reduced as temperature rises,
but that the total extension between -185° and 300° C. remains
noticeably constant. This clearly distinguishes aluminium crystals
from the hexagonal crystals, which, in general, experience sub-
 48. Progress and Termination of Glide. 157
stantial extension at higher temperatures. The increase in the
extension of aluminium crystals at 4000 and 5000 C. is followed at
6000 C. by a substantial reduction, due to the premature elimination
of primary octahedral glide by the emergence of a new (cubic) glide.
Recrystallization was observed in this case only at the highest
temperatures, and even then only at the apex of the extension cone
in the neck.!
The relationship between temperature and the yield-stress curve
based on primary octahedral
glide (glide along the system 7
which first becomes operative)
is shown in Fig. 128. While
at the start of glide the in-
fluence of temperature is ~5
only small, here, too, the -t
";'11-
effect of temperature on the "'"
increase of the shear stress .~
'" 3
becomes very marked as the ~
s
strain increases. '"
The influence of the test ~ JOOO
speed on shear hardening t1j 0
'I(}.
has been studied at room 500,0
6'O(JO,
temperature only (256). The o 20 6'0 80 100 1&0
yield-stress curve obtained
Glide stmin %
in dynamic stressing at a
FIG. I28.-Yield-Stress Curve of Al
23,000-fold speed (450 per Crystals in Relation to Temperature
cent. elongation per second) (257).
exhibits shear-stress values
which are approximately 16 per cent. higher than for the static
curve. A test which was carried out at an exceptionally reduced
speed of loading (test period: 234 days) is described in (258).
Particulars of the qualitative influence of time on glide are avail-
able also for IX-iron(259), tin (261) and bismuth (250). In all cases
an increase of the speed of testing leads to more pronounced harden-
ing and to a reduction of plasticity. This is particularly noticeable
1 The lattice position in which cubic glide undoubtedly starts enables us
to estimate its yield stress relatively to that of an octahedral system, as was
mentioned in Section 42. It is true that this will apply to extended crystals
only; the corresponding ratio for the intact initial crystal need not be the
same. This estimate is based on the fact that, in the end position [1l2]
towards which octahedral glide tends, cubic glide occurs, and leads to a new
end position [Ill]. Consequently, the shear stress in the cube plane in the
direction of a face diagonal exceeds that of an octahedral glide system by
only about 15 per cent.
158 Plasticity and Strength of Metal Crystals
in the case of the tin crystal, owing to the simultaneous operation
of several glide systems in rapid stretching. The same effect is
produced by a reduction of the test temperature to -185° C.
49. Crystal Recovery
In the previous section we dealt with the significance of tempera-
ture and time in the plasticity of crystals, and we gave particulars
of the yield-stress curves in relation to temperature and speed of
deformation. In so far as an influence exists, it takes the form of a
reduction of the hardening with rising temperature or diminishing
speed of deformation, for a given deformation. The experiments

J 11[
V
~ -j[
11J1J
glJ
.~ ~ 81J
-- IV
~ 5 71J.
8 oil
~ 198 ~ SIJ
'f(J E
JIl~
.?Il
11l •.
~-I
IJ ri', ~,,,78$'fI7fTr,:
1960 20 If{} 60 80 100 120
Tirne in seconds------,?- Tirne in rninutes
FIG. 129.-Flow Curves of a Sn Crystal FIG. 130.-Recovery Diagram of a
after Hardening and Recovery (261). Sn Crystal (250).
Between Tests III and IV the Unloaded
Crystal was Heated for 1 Minute at
60°C.

which we are about to describe, and which have led to the con-
ception of " crystal recovery", illustrate even more directly the
softening effect of temperature and time [(264), (266), (265)]. In
addition, they clearly reveal the close connection between these two
factors.
Fig. 129 contains flow curves of a tin crystal after repeated stretch-
ing at the same stress, and after a rest of 1 minute at 60° C. The
curves were recorded in a Polanyi filament-extension apparatus
(see Fig. 75). All extensions took place below the yield point of the
crystal. The initial reduction in the mean speed of flow reveals the
hardening which accompanies the deformation; while the weakening
effect of intermediate annealing is shown by the resulting increase in
 49. Crystal Recovery 159
flow. This weakening, which is not a result of recrystallization or
a change in orientation of the crystal, can be modified within a wide
range by changing the conditions of annealing. The measure of the
softening (e) is obtained ve. 100, where Vo
from e(%) = v'o -
- Ve Vo
represents the original flow speed, Ve the flow speed of the hardened
crystal, and v'o that of the recovered crystal. In the limiting case
of v'o = Ve there is no recovery; if, on the other hand, the original
flow speed is again reached (v'o = vo) recovery has been 100 per cent.
A three-dimensional diagram, based on such tests, which shows the
recovery of zinc crystals as a function of time and temperature, is
given in Fig. 130. Recovery can be considerable even at room
temperature over sufficiently long periods; it becomes perceptible
after 10 minutes and amounts to about 50 per cent. after 20 hours.
Analogous flow tests proved that bismuth crystals also possess a
marked capacity for recovery (250).
An even more striking demonstration of the softening effect of
time and temperature is shown in Fig. 131, which shows the exten-

1000

800

500 8

.... _-- ----- ._----

o o
o so 100 ,so lOO 0 '00 150

Extension % Extension %
a b
FIG. 131 (a) and (b).-Extension of a Zn Crystal; Tests Interrupted for
30-40 sec. in (a), and for 24 hours in (b), to Permit of Recovery.

sion curves of a zinc crystal. After each extension of 50 per cent.

the load was removed from the crystal for 30-40 seconds in the one
case, and for 24 hours in the other. The softening which occurred
during the interval is expressed in the reduction of the stress neces-
sary for further extension (from B to B'). The broken curves in the
diagrams represent the extension curves calculated theoretically on
the assumption of constant shear stress of the glide system (formula

•
160 Plasticity and Strength of Metal Crystals
43/1). It will be seen that the short pause resulted in only a slight
softening, whereas in the 24-hour interval the strongly increased
yield stress of the basis reverted to its original value. Therefore the
crystal recovered completely during this period. The measure of
recovery in these tests is given by the expression

0/) S
e ( /0 = Sl- '1100
S _ S' ,
1 0

So being the initial yield stress of the unstressed crystal (Point A),
SI the yield stress at the close of the first stage of extension at 50 per
cent. (Point B), and, finally, S' 1 the yield stress of the pre-worked
and recovered crystal (Point B').
The occurrence of recovery in aluminium crystals is clearly appar-
ent in creep tests (flow under constant load) at elevated temperature
(250° C.). In this case the extension of the crystals proceeds in
three different stages. Upon application of the load the crystal at
first flows rapidly, and in so doing forms numerous glide bands; then,
as hardening begins, there follows a stage of small deformation in
which only a few new glide bands occur, while finally, as a result
of crystal recovery, rapid flow occurs which leads to fracture
(258).
Tungsten crystals, too, exhibit very distinct recovery after suitable
heat treatment (at least 10 minutes at 2100° C.). This was revealed
by the considerable reduction in the ultimate tensile stress of
crystals that had hardened by drawing through diamond dies. The
original softness of the crystals was restored by this heat treatment,
indicating that here, too, the shear stress in the glide systems had
been reduced (266).
Although the direct experimental investigation of crystal recovery
has been limited to a few metals only, there can be no doubt that we
are faced here with a very general phenomenon which is technically
of great importance. In summing up it may be said that the
essential feature of crystal recovery is that, while preserving the
lattice, it continuously reduces, during heat treatment, the mechanical
properties of the work-hardened crystals approximately to the
initial values. Consequently, crystal recovery contrasts sharply with
recrystallization which involves the formation of new grains and
reduces the properties of the hardened metals down to their original
values in stages (cf. Section 65).
Hitherto the weakening effect of recovery has been discussed in
relation to the resistance of the crystals to plastic deformation; it
will be shown later that recovery influences also their tensile strength.
 50. Glide: Superimposition of Recovery upon Athermal Process Hi!

50. Glide Regarded as the Superimposition of Recovery upon an

Athermal Fundamental Process
In the previous sections the influence of temperature and time
on the progress of glide have been discussed. Thc main facts which
have been established, chiefly with hexagonal crystals, can be
summed up as follows: the critical shear stress which characterizes
the start of glide is only slightly dependent upon temperature (and
the speed of loading). There is no reason why it should not be
assumed that the shear stress in the glide system at absolute zero is
of the same order of magnitude as at elevated temperatures.
In marked contrast to this limited dependence on temperature of
the initial yield stress, the increase in the shear stress in the course
of glide (hardening) depends very strongly on the temperature, at
any rate within certain ranges. At very low temperatures both the
start and the slope of the yield-stress curve are practically independ-
ent of temperature. On raising the temperature a range is reached
in which the hardening is appreciably reduced as the temperature
increases; and finally the hardening curve, which now rises only
very slightly, again becomes almost independent of temperature.
The speed of deformation has a very similar influence on the
yield-stress curve: there is an insignificant effect at very low
temperatures and near the melting point, but a very strong influence
exists in the intermediate range.
The following physical description of crystal glide is suggested by
these facts (251). The basic process is athermal, i.e., independent
of time and temperature. It operates undisturbed at very low
temperatures, and finds expression in a characteristic function of the
material, the yield-stress curve, which indicates that the low yield
stress of the glide system of the unworked crystal greatly increases
with increasing deformation. Upon this basic process, which is
independent of temperature, there is imposed, as the temperature
rises, a thermal-softening process: c1'ystal recovery. In general,
therefore, two fundamentally distinct processes are simultaneously
involved in the deformation of a crystal by glide: the one concerns
the actual crystal and is structurally determined; the other is based
on the thermal vibrations of the lattice. As temperature rises, the
thermal recovery increasingly cancels out the hardening caused by
glide, leading ultimately in extreme cases to a deformation which is
free from hardening and which takes place at a constant shear stress
of the glide system.
From the above it will be immediately apparent that an increase
M

,.
162 Plasticity and Strength of Metal Crystals
of time, i.e., a reduction of the speed of testing, has the same effect
upon the softening process as an increase of temperature. This is a
plausible explanation of the influence which speed exerts on the
yield:-stress curve.

C. THE DYNAlVIICSOF TWINNING

51. The Condition of Twinning

Whereas glide could be characterized dynamically by the critical
shear stress and the yield-stress curve, similar quantitative principles
governing mechanical twinning are not yet available. As already
explained, the formation of a deformation twin represents, contrary
to glide, a process that is also macroscopically discontinuous.
Fig. 132a illustrates this by the extension curve of a cadmium
crystal (Polanyi filament-stretching apparatus), which clearly
reveals the formation of twin lamellre (accompanied by audible
crackling) resulting from discontinuous changes of load. The
twins themselves are usually narrow lamellre which often do not
traverse the entire crystal; and from these lamellre initially only
occasional small bands develop, which sometimes widen as the
stress increases, or are replaced by the formation of new twins .
.During the formation of a twin lamella a further increase of load is
unnecessary. It will suffice, therefore, for a dynamic characteriza-
tion of twinning if an account is given of the conditions which govern
its initiation. The reasons why we are unable at present to make
any statement regarding this" critical" state are to be sought on
the one hand in the very pronounced sensitivity of twinning to
inhomogeneities within the crystal, as a result of which the deforma-
tion twins always develop in a more or less extended stress range.
On the other hand, the attempt to formulate quantitatively the
conditions which govern twinning is greatly hampered by the
circumstance that, as a rule, the crystals also exhibit glide, in which
event, of course, twinning cannot be studied in the intact initial
state but only in an already hardened state. An example of this
will be found in the extension curves of variously oriented cadmium
crystals shown in Fig. 132, band c.
The few indications at present available for the initial law
originate from alternating-torsion tests. These tests, which were
carried out on zinc and bismuth crystals, would appear to indicate
that a single initial law based on a limiting stress, such as governs
glide, does not exist for mechanical twinning [(268), (269)]. It is
thought that twinning obeys an energy condition for which the path
180 Plasticity and Strength of Metal Crystals
removal of the stress, has been investigated on zinc and tungsten
crystals (288). The crystal wires were twisted by certain angles
(5-100 in the case of 10-cm.-Iong zinc crystals, 1800 for the 50-cm.-
long tungsten crystals) and then released. Whereas the tungsten
crystals immediately regained their initial position, the iinc crystals
retained a substantial permanent torsion. In both cases, however,
contrary to comparative tests carried out with polycrystals, the final
position was already reached 1 minute after removal of the load.
No further changes occurred during the investigation, which
extended over 36 hours, whereas in the case of poly crystalline wires
over the same period there was a subsequent reverse deformation
amounting to 40 divisions on the scale (corresponding to an angle of
torsion ~2°). These experiments show that the torsional after-
effect of metal single crystals, if it exists, is very much smaller than
with poly crystalline material.
Elastic hysteTesis means that the shape of a solid is dependent
not only upon the instantaneous value of the stress but also upon
the stress-strain history of the material. So far investigation of the
effect in single crystals has been limited to aluminium (291). Tensile
loads were applied to the crystals slowly and then slowly removed,
and the resulting changes in length were observed by means of the
Martens mirror apparatus. The stresses were kept within the range
0·028-1·43 kg./mm.2. The total extension after a large number of
load cycles amounted ultimately to 2·2 per cent. The results of the
tests can be summarized as follows: plastic extension already
occurred at the first application of load, even within the minimum
stress range of 0·028-0·144 kg./mm.2; therefore a range of a purely
elastic deformation was not observed. Repeated loading and un-
loading, after the initial occurrence of permanent deformation, led
ultimately to closed hysteresis loops similar to those observed with
polycrystals. The existence of these loops is not compatible with a
homogeneous elastic deformation of the lattice.
The Bauschinger effect consists in the observed fact that, after a
test bar has been deformed, its resistance to deformation (elastic
limit, yield point) is greater for further deformation in the same
direction than for deformation in the reverse direction (289). In
certain circumstances, even a weakening can be observed in the
reverse direction. For single crystals the Bauschinger effect was
studied on ()(-brass(72 per cent. copper). Cylindrical bars 9·2 mm.
in diameter were subjected to tension-compression tests, the changes
in the 10-mm. gauge length being observed by means of a Martens
mirror apparatus. The results obtained with the single crystal
 55. After-effect and Hysteresis. Bauschinger Effect 179
M. Polanyi (287) created the conception of strain strengthening
and contrasted it with shear hardening. He observed strain
strengthening on deformed tungsten and rock salt crystals (286).
Only with rock salt is the surface of fracture a smooth plane. The
strengthening of tungsten is assessed on the basis of the strength
related to the final section.
Finally, it may be mentioned that in deformed crystals strain
strengthening, like shear hardening, can be gradually removed by
crystal recovery. Figures relating to this phenomenon, for the
fracture of zinc crystals on the prismatic-plane type I, are contained
in Table XVII. After extension at room temperature, and before
fracturing at -185° Co, the crystals were heated for 2 minutes at
80° C. This produced a tensile weakening of 30-40 per cent. with-
out recrystallization of the crystal. Just as deformation produces
both strain hardening and strain strengthening, recovery leads to
both a softening and a weakening of the crystal (decrease of critical
shear and normal stress).

E. AFTER-EFFECT PHENOMENA AND CYCLIC STRESSING

Hitherto we have examined the behaviour of metal crystals under
substantial plastic deformation. We will now describe those closely
related phenomena which occur within a range of small deformation
upon removal of the load or change of the direction of stress. In
the first place it is a question of after-effect and hysteresis which,
although usually termed elastic processes, must no doubt be re-
garded essentially as plastic. The Bauschinger effect will also be
discussed in this context since it is probably due to the same causes.
The available material relating to these phenomena in metal crystals
is at present confined to a few experiments. A close enquiry into
this behaviour will, however, certainly contribute greatly to our
knowledge of crystal deformation and hardening.
Subsequently we shall describe investigations carried out on metal
crystals in regard to the technically very important property of
fatigue. In discussing these experiments we shall deal first with the
external crystallographic features of the crystals resulting from
cyclic stressing (deformation phenomena, crack formation), and then
with the changes which take place in the mechanical properties as
the stress increases.

55. After-effect and Hysteresis. Ba~tschinge1' Effect

The elastic afte1o-effect, which consists of an " after-shortening",
i.e., establishment of equilibrium length as a function of time after
178 Plasticity and Strength of Metal Crystals
cleavage planes are independent of the glide strain, as is more or
less the case at this temperature, then the limiting glide strain which
is achievable for crystals of various initial orientations is obtained
by intersection of the series of curves with the horizontal straight
lines at distance NB or N r from the abscissa axis. It will be seen
that where the initial positions of the basal plane are oblique, its
c!'itical normal stress (180 g./mm.2) is obtained only after consider-
able glide. Fracture would have to be preceded by a very sub-
&~antialdeformation. It has been found, however, that even after
re~atively small deformations the critical normal stress (1800g. Imm. 2)
for the prismatic plane is reached. These crystals fracture therefore
by cleavage along the prism plane.
Beyond an Xo angle of approximately 25° this prismatic cleavage is
then replaced by fracture along the basal plane. Whereas formerly
the glide strain increased with increasing xo' it now begins to
N
diminish. For a Xo angle corresponding to the equation tgXo = SO
o
the critical normal stress is already reached at the yield point, so
that for all Xo angles > arc tu ~ 0
(~58°) brittle fracture of the
o
crystals occurs.! The critical normal stress was here assumed con-
stant; however, even if it depended on the preceding deformation
the examination would follow a similar course.
Although the interpretation of the fracture process at low tempera-
tures as one of shear fracture is contradicted by the appearance of
the fracture surfaces of zinc crystals, it should be pointed out that
this view is also untenable for another reason. If the onset of
fracture resulted from the limiting shear stress, then this would
always have to operate (owing to the yield-stress curve which is
valid at low temperatures) upon reaching a certain point on this
curve, that is, at a constant limiting glide which is independent of
orientation. This does not agree with experimental results.
Apart from this case of the zinc crystal, which made possible an
investigation of strain strengthening of an operative glide plane (the
basal plane) and of a plane which crossed this (prismatic-plane type I),
there are no other examples of the strain strengthening of metal
crystals by glide. The question of the influence of mechanical twinning
upon the normal strength of cleavage planes is still quite obscure.
1 It should be noted that all curves in Fig. 143a have the straight line
N = k as asymptote. Whereas sooner or later fractlIre occurs in all orienta-
tions for values of the critical normal stress N <k, for N>k a fractme about
the plane under consideration is possible within only a restricted range of
orientation, and then only if there has been no previous glide.
 54. Influence of Previous Deformation 177
considerable basal glide at room temperature has a very pronounced
influence on the normal strength of the prismatic plane (determined
at -185° C. immediately after stretching). The varying amounts
of extension were achieved through the use of crystals of different
initialorientations. Table XVII gives the fracture.strength values
obtained, to which the critical normal strengths of the prism planes
are proportional owing to the very similar lattice position in stretched
crystals. The increase of the strain strengthening ofthe prism plane
with increasing cxtension is very clear.
From a knowledge of the behaviour of the critical normal stresses
for the basal and prismatic planes it is possible to understand tensile

2900

~ 'GOO

10

o 50 fOO '50 200 350 50 fOO '50 200 250

Glide s/min % Glide stmin %
(a) Basal plane. (b) Prism plane type 1.
FIG. 143 (a) and (b).-Theoretical Course of the Normal Stress when Extending
Zn Crystals of Various Orientations at -185° C. The Initial Angle
between Basal Plane and Longitudinal Direction is Indicated for the
Respective Curves.

tests with zinc crystals, and, in particular, the deformation which

precedes fractlu"e at those temperatures where fracture occurs in one
of the cleavage planes. In Fig. 143 is plotted the increase, during
the tensile tcst, of the normal stress on the basal plane and on the
prismatic-plane type I which lies most transversely, deduced from
the yicld-stress curve for crystals of various initial orientations at
-185° C.1 If the critical normal stresses (NjJ and Np for the two
1 NB was calculated with the aid of the relationship
NB = ~o t
ka _ (So = 126 g./mm." k = 400 g./mm.'),
a -I- cot g Xo
which is easily deduced from the formulre (40/2) and (43/3): while N p was
calculated from the equation Xp = cos' 30 (<Yo-I- ka) . (a -I- cot g xo)'
N
176 Plasticity and Strength of Metal Crystals
The effect of this deformation on the critical normal stress of the
basal plane is seen from Fig. 142. Although the results of the
individual tests are scattered, the following facts do at least emerge.
As was pointed out in Section 53, at -185° C. the influence of
preceding glide is only slight; at _82° C. there is a marked increase,
at -253° C. a pronounced decrease in the critical normal stress with
increasing glide, which in this case just exceeds 20 per cent. This
shows that, according to the temperature, the normal strength of
the glide plane can be increased or reduced by preceding glide.
The influence of strain hardening, due to deformation, on the
critical normal stress of cleavage planes (" strain strengthening")

TABLE XVII
Strain Strengthening and Recovery of Zinc Crystals (279)
-
Fracture strength at -185° C.

Extension, H,ecovery,
%. After recovery for %.
Immediately after
2 minutes at
extension,
80° C.,
kg·fmm.2•
kg·fmm.2•

50 2'6 - -
100 3·2 - -
160 5·5 3·4 38
300 7·3 - -
350 8·2 - -
380 8·4 6·2 26
400 9·6 5·4 44
500 9'8 - -

can therefore be negative, in which case it represents a weakening.

This contrasts remarkably with the effect of strain on the yield
stress of glide systems. In the latter case there is nearly always a
hardening (shown by the yield-stress curve) which increases with
decreasing temperature and finally reaches a limiting value (see
Section 48). If, at the yield point of zinc crystals, normal stresses
of about 500 g.jmm.2 on the basal plane at 20° C. do not lead to
fracture, this may perhaps be attributed to a strain strengthening
which is caused at this temperature by the slight extension which
occurs before the yield point is reached.
With the prismatic cleavage plane of zinc crystals as well as the
basal plane there appears at -185° C. to be no marked influence of
the basal glide which precedes fracture. For extensions between
11 and 48 per cent. the critical normal stresses scatter unsystematic-
ally between 1330 and 2130 g.jmm.2 (278). On the other hand,
 54. Influence of Previous Deformation 175
stresses. This is partly due to the difficulty of ensuring correGt
gripping of a cylindrical specimen when testing brittle cleavable
crystals. The magnitudes of the normal strengths of the cleavage

x
0

x
0 x

• 0

x
0

-
. 0
0
0
0

.
0 x x x
1---.
0
1----
0
eo
o.
'" 0
x-
•
x
X
\

1U 211 .JO qp 60 120 100 2110

Glide stmin % Glide strain %
(a) at -2530 C. (c) at _820 C.

.
• x X
0

·ff . ",.. . . . .
OR
00

0
. .0

x .
60 120 100 2PP 2qp 260
Glide strain %
(b) at -1850 C.
FIG. 142 (a)-(c).-Influence of Previous Glide on the Critical Normal
Stress on the Basal Plane of Zn Crystals at Various Temperatures (281).

planes are of the same order as those of the shear strengths of the
principal glide systems (Table IX).
Further tests are necessary if the significance of the critical normal
stress in a quantitative description of cleavability is to be elucidated.

54. Influence of Previous Deformation upon the Critical Normal

Stress. "Strain Strengthening"
It was pointed out in the previous section that the fracture of
metal crystals on cleavage planes is often preceded by plastic
deformation. This is specially obvious in the case of zinc crystals.
174 Plasticity and Strength of Metal Crystals
the cleavage plane. The value obtained for the critical normal stress
of the rhombohedral plane, which is more than double that of the
(Ill) plane, can therefore be taken only as an approximation.
Only scanty experimental information is so far available regarding
the dependence of the critical normal stress on temperature, and in
any case we have seen that such dependence can be only slight.
Thus the tensile strengths for bismuth at 20° and -80° C. can be
represented by the same value of the critical normal stress of the
basal plane (Fig. 137). And even for the zinc crystal there appear
to be no noticeable differences in the basal normal stress within the
range -80° to -253° C. (Section 54).

TABLEXVI
Critical Normal Stress of Metal Crystals
O· _

Critical
Method of Tempera- normal
Cleavage Litera-
Metal. producing ture, stress at
plane. o C.
ture.
the crystal. fracture,
kg./mm.2.

Zinc (0·03% Cd) . (OOQl)} 0'18-0'20 (278)

(1010) -185
Drawn from 1·80 (281)
Zinc + 0'13% Cd the melt. (0001) -185 0·30
Zinc + 0'53% Cd (0001) -185 1·20 } (280)

Bismuth Drawn from

the melt
(Ill) +
-
20
80
)
f
0·324 (284)
and soli- (ll!)} 0·29
dified in (ll!) + 20
0·69 } (285)
Antimony vacuo (111) + 20 0·66

Tellurium Solidified
slowly
(1010) + 20 0·43 (283)

Scarcely anything is known of the influence of metals in solid

solution on the critical normal stress. In the case of zinc crystals
alloyed with cadmium there appears to be quite a substantial
increase in the fracture strength (280).
A summary of the critical normal stresses so far measured is
given in Table XVI. In addition to the examples already mentioned,
zinc (fracture along basal and prismatic planes, alloying effect),
bismuth (fracture along basal and rhombohedral planes, influence
of temperature), and tellurium, the table contains an approximate
value for the critical normal stress of the rhombohedral plane of
antimony. It should be particularly noted that the values for the
normal stresses are less reliable than those for the critical shear
 53. Brittle Fracture: Sohncke's Normal Stress Law 173
glide, constancy of the shear stress can at present be described only
as much more probable than constancy of strain, in the case of
brittle fracture of metal crystals the balance is clearly in favour of
the constant-stress law.
The zinc crystal affords also an opportunity for determining the
critical normal stress ofthe second-best cleavage plane, the prismatic-
plane type I, which appears as cleavage plane when the basal
positions are very oblique. At -185° C. a value is obtained for
this which is about ten times greater than for the basal plane; it is
measured, of course, after slight basal glide. There is an order-of-
magnitude preference for the basal plane, not only for glide but also
for cleavage fracture.
Taking into account the critical normal stresses of basal and
prismatic-plane type I it is now possible to calculate generally the

a
FIG. 141 (a) and (b).-(a) Tensile-strength Surface, and (b) Fracture
Surface, of the Zn Crystal at -185° C. (278).

dependence of the tensile strength upon orientation. The result is

represented by the surface illustrated in Fig. 141a, which gives the
tensile strength corresponding to various orientations and clearly
reveals the strong influence of crystal orientation. A surface repre-
senting the tendency to fracture in a wider sense is shown in Fig.
141b. It connects the initial orientation of the crystal with its
tensile strength with respect to brittle cleavage along the basal or
prism planes, and (in far the greater part of the field of orientation)
with its yield point with reference to extension by basal glide. This
is given by the constant critical value of the shear stress in the glide
system (Section 40).
For bismuth, too, the second-best cleavage plane-the rhombo-
hedral plane (lll)-could be characterized quantitatively by means
of its critical normal stress (285). It is true that, with the orienta-
tions investigated, mechanical twinning always started before
fracture occurred, as could be detected from narrow striations on
172 Plasticity and Strength of Metal Crystals
occurs (obtained from experiments) does not usually apply for the
un deformed initial state. It may depend very largely on the extent
of previous deformation, as will be explained in greater detail in
Section 54. At -185° C., however, this dependence was found to
be only slight, so that the tests at this temperature can be discussed
here (278). Fig. 140 repre-
3,0 sents the tensile strength
as a function of the angle
Z,5 of the cleavage plane.
The observations again
agree satisfactorily with
the curve calculated on
the assumption of the
normal stress law. In
particular, these tests with
zinc reveal substantial
disagreement between the
results of the tests, and
o a fracture condition of
'" 10

.l
S -0.5
constant normal dilatation
perpendicular to the cleav-
-- •••...•. age plane. A discontin-
-1,0
uous line in Fig. 140
indicates the course of
-1,5 the tensile strength re-
FIG. 140.-Tensile Strength of Zn Crystals sulting from this assump-
at _185° C. as a Function of Orientation
[(278), (282)]. tion. Agreement between
the two statements is
constant normal stresses: a = si~; x; assumed for the trans-
---: constant normal dilatation: verse position of the
a =
basal cleavage plane. It
811 COS X + 833 sm X
2 £2 . 2
will be seen that con-
stancy of the normal
dilatation would demand a very much more pronounced depen-
dence of the tensile strength upon orientation. For a position

angle of ~26°(tgx. = J- ~13)33

an infinitely great tensile strength
is obtained (dilatation resulting from longitudinal extension, and
reduction of area resulting from transverse contraction of the test
bar, are here equal to each other) and for a still more oblique position
of the cleavage plane normal dilatation becomes possible only under
compressive stress. Whereas in the case of the initial condition for

•
 53. Brittle Fracture: Sohncke's Normal Stress Law 171
The tensile strength increases, therefore, as the direction of tension
approaches the hexagonal axis. Down to an angle of inclination of
29° for the most transverse (1010) plane the normal stress law
continues to apply (Fig. 138). If the position of the cleavage plane
becomes still more inclined, the strength of the materiall is exceeded
before the critical normal stress is reached. In this range of orienta-

FlG. 139.-Tensile Strength Smface of the

Te Crystal (283).

tion the tensile strength remains noticeably constant at approx.

1·8 kg.jmm.2 A spatial representation of the dependence of the
tensile strength of tellurium crystals upon the orientation is given
by the surface illustrated in Fig. 139.
Zinc crystals, which at 20° C. and at higher temperatures do not
as a rule fracture along smooth crystal surfaces, exhibit at -80° C.
(and at lower temperatures) smooth cleavage fractures along the
basal plane. This fracture is mostly preceded by glide along the
basal plane, which at -185° C. can amount to a glide of approx.
200 per cent. Therefore the basal normal stress at which fracture
l Tmnslator's footnote. \Vith respect to fractul'e processes other than
cleavage along the (l010) plane.
170 Plasticity and Strength of Metal Crystals
which functions as cleavage plane (basal plane in the hexagonal
representation). The curve calculated according to equation (53/1)
on the assumption of a constant normal stress on the cleavage plane,
agrees with the observations. Since the (Ill) plane is also the best
glide plane of the bismuth crystal, the lattice positions which at
20° C. separate the area of brittle fracture from that of incipient

4.2

2,8
\
0
\
0

0
x
x

x
~x
\
..
0
x
0
~ I\x
x
x v
.,
x

X
\
"x
~
~-........'-
':,<

o 10° 20° 30° 'fOo 50° 60° 70° 80°

Xt
90"

FIG. 138.-Tensile Strength of Te Crystals as a Function of Orientation (283).

Xl angle between tensile direction and the most transverse prism plane
type 1.

glide can be calculated from the critical shear and normal stress of
this plane. From the condition .
S N sin X N
er = sin X . cos ), = sin2 X the relation cos A = 11
is obtained for the boundary of the areas of glide and of fracture.
With the aid of the critical stress values for bismuth we arrive at a
limiting angle of 55° 40', for X = A. Where the initial position of
the (111) plane is more transverse, it is the critical normal stress that
will be reached first, while with more obliquely oriented crystals it
will be the critical shear stress.
Results similar to those with bismuth were obtained with tellurium
crystals, which exhibited tensile-strength differences of 1 : 4. The
cleavage plane here is one of the (1010) planes (prism planes type I).
 53. Brittle Fracture: Sohncke's Normal Stress Law 169
(see Fig. 136). This fracture either occurs without any previous
perceptible deformation (brittle crystals), or it terminates a more or
less large deformation the magnitude of which substantially affects
the value of the tensile strength.
Nothing is yet known regarding the period of time covered by the
process of rupture along cleavage planes. Although as a rule
fracture occurs rapidly, it is, nevertheless, certain that the speed of
the operation is finite (see Section 69). Markings on the cleavage
planes and the occasional presence of flaws in the crystal fragments
indicate that the process of fracture often consists in the gradual
deepening of cracks.

53. Conditions Covering Brittle Fracture. Sohncke' s Normal

Stress Law
Tensile tests carried out on rock-salt crystals of various orientation
led Sohncke to formulate the Normal Stress Law (277). According
to this law these crystals break
when a certain critical normal 1500

.\
stress has been reached in the
1'IQ0
cubic-cleavage plane which is
invariably the surface of frac-
ture. The fracture strength 1 1200 '"
\.
therefore depends strongly upon <:i,
.~
1000
,\
the angle (x) between cleavage
plane and the direction of the
800 .
tension. If N is the critical 800
\.
normal stress, equation (40/2) ~ , 0

gives for the fracture strength (J , .."..

,, ~)

the expression (J = .~ (53/1). 200

~:
sm X 1:,'
,«'
(For further particulars of ex- ~. M' W· ~. ~. m' ~.
periments with rock salt see Angle between rod axis and cleavage
plane
Chapter VII.)
Systematic experiments with FIG. 137.-Tensile Strength of Bi
Crystals as a Function of Orientation
metal crystals have shown that (284).
here, too, the occurrence of brittle X = at - 80· C., 0 = at 20· C.
fracture along smooth cleavage
planes can readily be brought within the scope of the normal stress
law. For example, bismuth crystals fracture in a brittle fashion at low
temperatures and also at room temperature within a certain orienta-
tion range. Fig. 137 shows the extent to which the fracture strength
of bismuth crystals depends upon the position of the (Ill) plane
 168 Plasticity and Strength of Metal Crystals

(a) a-Fe
(001).

(b) Zn
(0001).

. - . -

(c) Bi
(Ill).
'..

-
_

- I

(d) Sb
(Ill).

(e) Te
(1010).

FIG. 136 (a)-(e).-Fracture (Cleavage) Planes of Metal Crysta.ls.

 52. Mutual Influence of Twinning and Glide 167
due to primary hardening by alloying, or to a substantial hardening
of a latent glide system, in the present case it is the pronounced
hardening of the basal plane due to twinning that is responsible for
the reduction of its capacity for further hardening.
Scarcely any information is yet available on the question of the
influence of mechanical twinning on further twinning-that is, whether
a deformation twin or the initial crystal is more resistant to
mechanical twinning. The answer depends on whether the plasticity
which can be realized through repeated twinning is parallel or
opposed to the direction of stress. It is also necessary to distinguish
between the untwinning of a deformation twin and the development
of a secondary twin which is distinct from the original crystal.
Tests carried out on zinc sheets (274) show that it is possible, by
bending backwards, to cause previously produced deformation
twins to disappear; in this case, therefore, untwinning proceeds
more easily than further twinning in the parent crystal. Similar
observations have been made with antimony crystals (275). A
more quantitative investigation of these relationships confirmed this
result (276). In the course of an alternating bend test carried out
on coarse-grained zinc sheets it was found possible to produce, and
untwin, the same twin lamellre as many as twenty times. Besides
sudden broadening (or narrowmg), in the course of these experi-
ments a phenomenon was observed which, under the microscope,
appeared to be a continuous change of the width of twin lamellre
with increasing (or again decreasing) bending moment. The
moment required for the complete untwinning of a lamella is greater
than that needed for its production. If the same lamella is repeatedly
produced, weakening will be observed in so far as the required
moment will diminish steadily. Consequently, if the test is repeated
with equal bending moments, broader lamellre will be produced.
On the other hand, the bending moments required for untwinning
appear to increase as the test is repeated.
This chapter shows that we have still a very great deal to learn
about the dynamics of mechanical twinning.

D. FRACTURE ALONG CRYSTALLOGRAPHIC PLANES

Hitherto we have been concerned with crystal deformation pro-

cesses, a consideration of which included glide and shear fractures.
In the present section we shall discuss that type of rupture which
consists of a fracture of the crystal along plane surfaces which
coincide invariably with one of the cleavage planes listed in Table VI
166 Plasticity and Strength of !fetal Crystals
it is transverse. Consequently, each of the three most obvious
assumptions: i.e., that either the true tensile stress, or the shear
stress, or the normal stress in the twinning system, is responsible for
the lower limit of the stress range in question, is compatible quali-
tatively with the experimental data relative to the width of the
stress range that leads to twinning. The experiments show further,
however, that the range within which twinning occurs decreases
with increasing Xo angle much more rapidly than would follow from
either of the three assumptions. If the effect is attributed to an
increase of the lower stress limit, it is then possible to say that
twinning is made more difficult by previous glide.
Only a few details are as yet
c available regarding the converse
.--- ox case of the influence of twinning
on glide. An examination of zinc
and cadmium crystals suggests that
B
the shear stress of the basal plane
experiences a considerable discon-
A tinuous increase as a result of
o twinning [(271), (272), (270), cf.
Glide stmin %---+-
also (273)]. This shear hardening
FIG. 135.-Sketch showing the
Behaviour of the Resolved Shear
by twinning amounts at room tem-
Stress on the Basal Plane in the perature to about 100 per cent.
Course of Primary Glide (AB), for zinc crystals, and about 200
Mechanical Twinning (BC) and
Secondary Glide (CD) of Zn per cent. for cadmium crystals. It
Crystals (272). is true that the resultant twin
bands, especially in the case of
cadmium, are usually very thin, so that it would appear possible
that the mechanical properties are influenced by the size of the
crystal investigated. Subsequent glide in the twin is accompanied
by only a slight shear hardening on the basal plane, frequently even
by a reduction of the shear stress (a softening during secondary
basal glide occurred in nine out of twelve cases with zinc crystals,
and in four out of twenty-three cases with cadmium crystals). A
diagram showing the behaviour of the resolved shear stress of the
basal plane of a zinc crystal is given in Fig. 135.
The fact that secondary glide on the basal plane is accompanied
by only slight hardening (or even by some softening) furnishes a
further example of the limitation of the hardening capacity of
crystals. Analogous examples have already been noted in the glide
of solid solutions (Section 46). Whereas in the case of solid solu-
tions the small extent of the further increase of the yield stress was
 52. Mutual Influence of Twinning and Glide 165
maximum load if the position of the basal plane is rather transverse
and the extension curve consequently flat.
This phenomenon can perhaps be explained as follows: if, taking
as a basis the yield-stress curve which is valid for basal glide, we
calculate the extension curves in the true tensile stress-strain
diagram, the result shown for cadmium in Fig. 134 is obtained.!
If allowance is made for the change in orientation which aCCOID-

woo
If
'1000

.1500
/20
j
/0'

/ / 7
1/5°
50'

I / 17
I 1/ }o'
/ / / /V
/ / /V
/ / / /
/
/
?'/
'100
/j / ~
./ ../ ../~
~
~
° 50 100 150 zoo Z5f} 300 350 '100 'I5Q
Extension %
FIG. 134.-CalcuJated True Tensile Stress-Strain
Curves for the Basal Glide of Cd Crystals of
Various Orientations.

panies extension, it is also possible to deduce from these curves the

behaviour of the shear and normal stress in the preferred twinning
system. The curves obtained for these stresses exhibit in their
general course no substantial difference from the true stress curves.
In all three cases the high values are obtained with much smaller
extensions if the initial position of the basal plane is oblique than if
1 The true stress at,. (referred to the current cross·section) is obtained from
the" conventional" stress a referred to the initial section by multiplying by
d = 9JJ. In order to obtain the equation of the true stress curves, the equation
q
(43j3) must be multiplied by d.
L64 Plasticity and Strength of Metal Crystals
of the atoms from their starting points to their fuIal position is also
of importance. One fact does emerge, however, from the scanty
material so far available regarding mechanical twinning of metal
crystals, namely, that as with glide (see Section 42), the shear of
the two neighbouring lattice planes is not essentially elastic. The
elastic shear value which is reached at the onset of twinning is many
orders of magnitude below that of the shear resulting from twinning.
Let the cadmium crystal serve as an example: the reciprocal shear
modulus in the twinning system is certainly not greater than
60 X 10-13 cm.2/dyne, the maximum value for l/G. The shear
stress needed to produce twin lamelhe in the twinning system is also
substantially smaller than 500 g./mm.2 ~ 5 X 107 dyne/cm.2 (270),
so that the elastic shear attained at the moment of twinning is

FIG. 133.-Fe Crystal Extended at -185° C., Showing Deformation

Twins. ]{l = (112) (269a).

certainly smaller than 3 X 10-4 against SOd = 0·175. Differences

of the same order occur with other metals which form deformation
twins.
Temperature appears to influence twinning even less than it
affects the start of glide, for twinning is superseded by glide at high
temperatures (Zn, Cd), but occurs preferentially a,t low temperatures
(oc-Fe; Fig. 133).

52. Mutual Influence of Twinning and Glide

The width of the stress range in which mechanical twinning can
occur in hexagonal crystals under static tensile stress depends to a
very marked degree upon the orientation of the crystals, or rather
upon the shape of the extension curve as determined by the orienta-
tion of the glide elements. This can be seen from Fig. 132, band c,
which shows that twinning starts early where the initial position of
the basal plane is very oblique, and the extension curve therefore
steep, but that it is restricted to a very narrow range just below the
 51. The Condition of Twinning 163

1JOO

1200

~ 1100

] 1000
I
.~ ~I
.,"
<.l
!JOO

~Cl 800
ts
~ )
:~
.~
~
700

600
1/ X-7 r
A=1 °3
\~

500
I
/
11100 If/'
::e ~
Cl 600 .•.::e JOO
r:i
!. IJIJO ~ 200
.,,;, ' . Crystal No. 35 ~
•~ ZOO 100
.,.,
~ o 6 12 18 2'1 30 o 10 20 JO 'la 50 tiO
ji}xtension % Extension %
(a) Discontinuous change of (b) Twinning after substantial glide.
load in the formation of deform-
ation twins.

50 100 150 ZOO 250 JOO J50 'lOO /J50

Extension «J~

(c) Twinning after substantial glide.

FIG. 132 (a)-(c).-8tress-8train Curves of Cd Crystals (270).