R. J. D A V I S
In the use of this method as a setting method the
main difficulty arises in selecting the reflexions to be
brought into coincidence. This does not arise with irregular fragments of known unit cell (e.g. ground
spheres), and the method may then prove superior to
others. For example a colleague setting an irregular
crystal of (?) face-centred cubic ferritungstite to [100]
repeatedly converged on [111] by trial oscillation photographs. We had no difficulty in bringing two pairs
of 111 reflexions into off-equatorial coincidence by this
method, thus setting to the required axis. In some
similar cases it may well happen that the predicted
position of A in Fig. 1 lies outside the range of the arcs.
The prediction will then be rather inaccurate owing to
departures from the approximations above, but should
suffice to indicate how the crystal should be remounted
for a further triplet of trial rotation photographs. The
indication of azimuth is likely to be more accurate than
that of the inclination. It may be preferable to locate
another known axis, accessible to the arcs, and derive
the inclination from this.
Most workers use trial oscillation (or Laue) photographs for normal crystal setting, when using an irregular fragment. It is well known that recognizing
badly misset layer lines is a knack, to be acquired by
practice. The writer derived this method for general
use before acquiring this knack, to study serendibite
(Prior & Coom~irasw~imy, 1903; cf. Pertzev & Nikitina, 1959) available as type material in the form of
rough irregular fragments thought to be triclinic. The
273
above method was triumphantly successful at first attempt, setting a crystal on to an axis displaying equatorial symmetry on even-order, but not on odd-order
layer lines, by using off-equatorial coincidences. The
axis proved to be [122] of the reduced all-acute triclinic cell.
Subsequent use suggests that this success was partly
accidental. The difficulty lies in properly selecting the
reflexions to be brought into coincidence. Choice of
off-equatorial coincidences if possible will favour setting to a symmetry axis if any exists. It seems likely
that extending Fig. 1 to consider more than two reflexions, thus obtaining a consensus, offers best chance
of success. The method used in these circumstances is
simple and offers as much chance of quick success as
trial oscillation photographs, but for general setting of
irregular fragments, the method of Brooker & Nuffield, while more elaborate, is more certain. However,
where the unit cell of the irregular fragment is known,
the present method offers many advantages over other
recommended photographic methods.
m
References
BROOKER, E. J. t~ NUFFIELD, E. W. (1966). Acta Cryst. 20,
496.
PERTZEV, N. N. t~ NIKITINA, I. B. (1959). Mere. All-Union
Min. Soc. 88, 169. c f Miner. Abstracts (1959), 14, 274.
PRIOR, G. T. & COOMARASWAMY,A. K. (1903). Miner. Mag.
13, 224.
ROOF, R. B., JR (1955). Acta Cryst. 8, 434.
Acta Cryst. (1968). A24, 273
The Decomposition of an Anisotropic Elastic Tensor
BY YIH-O T u
Thomas J. Watson Research Center, Yorktown Heights, New York, U.S.A.
(Received 30 March 1967)
The classical theory of invariants asserts that there exists a finite integrity basis whose elements are
polynomials of strain components and are invariant under the group of transformation defining each
symmetry class of a crystal. By constructing a strain energy function made up of the elements of an
integrity basis for a certain symmetry class, we derive a tensor basis which spans the space of elastic
constants for crystals of this symmetry class. Introducing systematically new elements of the integrity
basis into the construction of the strain energy function, we construct five hierarchies of orthonormal
tensor bases which span the space of the second-order elastic constants of all crystal systems. Any
elastic tensor of rank four possessing certain crystallographic symmetry may be decomposed into a
sum of tensors of increasing symmetry. From this representation of an anisotropic elastic tensor, the
tensor of any given symmetry, not only the isotropic one, nearest the given tensor can be read off
immediately. Bases which span the space of elastic constants of orders higher than the second may
be computed in a similar manner. Such computations can be carried out by a computer. A FORMAC
program of 7090/94 IBSYS has been written to obtain the elastic constants of the second and the
third order for each class of a crystal.
1. Introduction
In an investigation of the physical properties of an
anisotropic body, one sometimes begins with the cor-
responding properties of an isotropic body having the
same geometry and proximal constitutive physical relations, and for which some knowledge may be obtainable with relative ease. It may then be possible to
274
THE DECOMPOSITION
OF AN A N I S O T R O P I C E L A S T I C T E N S O R
determine properties of the anisotropic body by the
method of perturbation. Ways of defining and constructing the isotropic elastic tensor 'nearest' a given
elastic tensor have been presented by Gazis, Tadjbakhsh & Toupin (1963). In this paper we construct
orthonormal bases of tensors of rank four which span
the space of elastic tensors possessing certain pointgroup symmetry of a crystal. The orientation of the
Cartesian coordinates (XbXz, X3) conforms to the conventions in I R E Standards on Piezoelectric Crystals
(Institute of Radio Engineers, 1949); namely,
Monoclinic
Tetragonal, trigonal and hexagonal
Orthorhombic and cubic
xzlly,
x3[lz, x~l[x,
XII[X,Xz[[y,X3[lz,
where (x,y,z) are the natural axes of symmetry of the
crystal.
General procedures for the construction of linearly
independent tensors of rank 2N, N = 2, 3 , . . . possessing
given symmetry, are presented in § 2. By means of these
tensors, orthonormal bases of tensors of given symmetry can be constructed and are given in § 3. The
elements of the orthonormal bases of tensors of rank
four are exhibited in both tensor and abbreviated
matrix indices. The correspondence between the tensor
indices and matrix indices is as follows:
Tensor indices: 11 22 33 23,32 31,13 12,21
Matrix indices: 1 2 3 4
5
6.
For the elastic stiffness tensor, we set
civet = CMN (i,j, k, l = 1,2, 3 and M, N = 1 , 2 , . . . , 6).
However, for the elastic compliance tensor s~j~t, factors
of 2 and 4 are introduced as follows:
SiJlcl m S M N when both M and N are 1, 2 or 3,
2sijel = SMN when either M or N is 4, 5 or 6,
4StjgZ=SMN when both M and N are 4, 5 or 6.
Consequently, the matrix elements in § 3 must be multiplied by the factors 2 and 4 according to the above
schemes.
To illustrate the method of decomposition, in § 4,
the elastic stiffness tensor of quartz is decomposed into
a sum of tensors of increasing symmetry. From this
representation of an anisotropic elastic tensor, the
tensor of any given symmetry, not only the isotropic
one, nearest the given tensor can be read off immediately.
2. Invariant tensors
When a perfectly elastic crystal, initially stress-free, is
deformed either isothermally or adiabatically to a final
stressed state, the strain energy if" is a function of state.
Furthermore, IT" is invariant under the finite group of
transformations, G, which defines the symmetry class
of the crystal. For the purposes of this paper, it is assumed that if" is expressible as a polynomial in the
strain components Eij. Accordingly, if I la~jll is an element of G, and E~j=a~kajtEkz,* then it is required that
I~(E~j) = I~(E~),
(1)
for every Ilaijll in G and all values of E~. A classical
result of the theory of invariants (Weyl, 1946) is the
existence of a finite integrity basis { l , , I 2 , . . . , I x } such
that every polynomial function satisfying (1) is expressible as a polynomial in the elements of the integrity
basis. Each element In of an integrity basis is itself a
polynomial in E~j satisfying (1). For the case of an
isotropic material, for example, an integrity basis has
three elements. They are polynomials of the first, second and third degree in Eij, respectively, and may be
chosen as follows:
I = t r (gij)=Eii=E11 + E22+ E33
II=½[EuEz-Ei:Ei:]
=(Ea,Ez2+E22E33+ E33E,,)-(E~2+E~3+E~,)
Ill =det (Eij)--~[E~IEj:Ek~- 3EuEjkE~j
+2E~:E:~Ekd
= ElxE22E33 + 2E12E23E3x
-
(2)
(ExlE23
2 + EzzE3,
2
+ fzzEh)
Smith & Rivlin (1958) have determined the integrity
basis for invariant functions of Eij for each of the crystallographic point groups. Their results are summarized
in Tables 1 and 2 for the purposes of the presentation
of this paper. Table 1 gives the number of elements
of an integrity-basis for each crystal class together with
the number of the second- and third-order elastic constants. Table 2 gives some additional elements of polynomials of the first, second and third degree which are
linearly independent of I, II and III of equations (2)
as well as among themselves, together with the symmetry classes for which it is an invariant.
The second-order, third-order,..., elastic constants
are tensors of ranks four, s i x , . . . , etc., and are defined
and denoted by
cijkz
1,
=
~Ei~E~tJ
f
~3~"
IIE~JlI=II011 '
1
(3)
C~j~mn= t~E~j~E~Er~nI IIE~JII=II011 '
. . . . . , etc.
These quantities are symmetric in the following pairs
of indices, (i,j), (k,l), (m,n) and (ij, kl), (ij, mn), (kl, mn).
In particular, c~jk~ is called the elastic stiffness tensor.
By choosing elements from Table 2 in addition to elements of (2), we shall form strain energy functions IT"
as polynomials of second and third degree in Eij. Since
polynomials of degree greater than three in the integrity
basis do not contribute to the elastic tensors of ranks
* The summation convention on the double indices of
tensors is understood throughout this paper.
YIH-O
TU
275
T a b l e 1. Number o f elastic constants and number o f elements in an integrity basis f o r each crystal class
N u m b e r of elements in an
integrity basis
Crystal
systems
Class symbols
Sch6nflies International
Triclinic
~ CI
t C,
Monoclinic
Orthorhombic
Tetragonal
1
T
Hexagonal
Cubic
1st
degree
6
6
2nd
degree
0
0
3rd
degree
0
0
4th
degree
0
0
5th
degree
0
0
6th
degree
0
0
2nd
order
21
21
3rd
order
56
56
C2
Cs
C2h
2
m
2/m
4
4
4
3
3
3
0
0
0
0
0
0
0
0
0
0
0
0
13
13
13
32
32
32
D2
222
C2v
D2h
C4
$4
C41~
D4
C4h
D2a
D4n
C3
mm2
mmm
4
Z~
4/m
422
4mm
~2m
4/mmm
3
3
3
3
3
3
3
1
1
1
0
0
0
0
0
0
0
0
0
9
9
9
20
20
20
2
2
2
2
2
2
2
4
4
4
3
3
3
3
4
4
4
2
2
2
2
2
2
2
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
7
7
7
6
6
6
6
16
16
16
12
12
12
12
2
2
2
2
2
4
4
3
3
3
8
8
4
4
4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
7
7
6
6
6
20
20
14
14
14
2
2
2
2
2
2
2
2
2
2
2
2
2
2
4
4
4
2
2
2
2
2
2
2
1
1
1
1
2
2
2
1
1
1
1
2
2
2
1
1
1
1
5
5
5
5
5
5
5
12
12
12
10
10
10
10
1
1
1
1
1
2
2
2
2
2
5
5
3
3
3
3
3
2
2
2
2
2
1
1
1
1
1
0
0
0
3
3
3
3
3
8
8
6
6
6
1
1
1
0
0
0
2
3
C3i
Trigonal
N u m b e r of
elastic constants
D3
Car
32
3m
D3a
C6
]~m
6
C 3t~
"~
C6h
6/m
D6
622
C6v
D3~
D6n
6mm
~m2
6/mmm
T
Tn
O
23
m3
432
Ta
Oh
~3m
m3m
Isotropic
T a b l e 2. First, second and third degree anisotropic polynomials and their &variant properties
Polynomials
1st
degree
2nd
degree
E33
Ell
E31
El22 + E232 + E3I 2
E232 + E312
E12(E1 t - E22)
E23(E11 - E22) + 2E12E31
E3 I(E11 -- E22) -- 2E12E23
E232
EIEE23
Ell E332+ E22Ell z + E33E222
E11E312 + E22Elz 2 + E33E232
EllE22E33
E12E23E3t
E23E31(Ell - E22)
E12(E232 - E312)
3rd
degree
E23(E232 - 3E312)
E23[(E11 + E22) 2 + 4(E12 z - E22z)] + 8EI1E12E31
E31 (E312 -- 3E232)
E31[(E11 + E22) 2 + 4(Elz 2 -- E222)] - 8EllElzEz3
E23E31(Ell - E22) + E12(E232 --E312)
3EI2(E11 -- Ezz) z - 4Elz 3
E12(E312 -- E232) + E23E31(E22 - E33)
E11[(E11 + 3E22) z -- 12E122]
Invariance
Tetragonal, trigonal, hexagonal, orthorhombic, monoclinic
Orthorhombic, monoclinic
Monoclinic
Tetragonal, cubic
Tetragonal, trigonal, hexagonal, orthorhombic, monoclinic
Tetragonal (classes 4, ~, 4/m)
Trigonal
Trigonal (classes 3, ~)
Orthorhombic
Monoclinic
Cubic (classes 23, m3)
Cubic (classes 23, m3)
Cubic
Cubic, tetragonal
Tetragonal (classes 4, ~, 4/m)
Tetragonal (classes 4, ~, 4/m)
Trigonal
Trigonal
Trigonal (classes 3, 3)
Trigonal (classes 3, 3)
Trigonal (classes 3, 5)
Trigonal (classes 3, 3), hexagonal (classes 6, ~;, 6/m)
Hexagonal (classes 6, ~, 6/m)
Hexagonal, trigonal
276
THE DECOMPOSITION
OF AN A N I S O T R O P I C
four and six, as is evidenced from (3), such truncated
polynomials of ~ are sufficient for the determination
of the second- and third-order elastic constants according to (3). In the following, we shall show the construction of bases in the space of the second-order
elastic constants.
To begin with the isotropic integrity basis (2), the
second-order elastic constants are obtained from terms
of 12 and II of (2) in the strain energy if'. By carrying
out the differentiations of (3), one obtains the following
two isotropic tensors of rank four,
fl~a = 6,~g~ + & ~ j ~ . .
(4)
We now introduce the element (E 2~2+ E23
2 + E321) of
Table 2 into the construction of W. Differentiation of
the term involving this new element in if" gives the following tensor of rank four which is invariant under
the group of transformations characterizing either a
cubic or a tetragonal crystal,
~i]/Cl = ((~ 1~3"2~- ~2(~Jl)(~/C1~/2 -1t- ~/¢2~/1)
-1- (~i2~j3 -1- ~3~3"2)(0/C2~/3 21-g/C36/2)
Consequently, the three elements (e,fl, re) of (4) and (5)
constitute a tensor basis for the space of the secondorder elastic constants of a cubic crystal. With the additional introduction of the elements E33 and ( E ~ +
E]~) of Table 2, the three additional second degree
terms E2s, E33(Ell + gz2 + E33) and (EzZs+ E2~) in if"
produce three more tensors of rank four in addition
to (~,/~, ~); they are
& m = &j&c3&3+ &3J~3Jet
+ (&~O~3+ &3O~l)(&~&3+ &3&l) •
= &~6kt
/
(6)
ELASTIC TENSOR
Therefore, the six elements (e, fl, z~,),, g, e) of (4), (5) and
(6) form a basis for the space of the second-order elastic
constants of crystals of classes 422, 4ram, 42m and
4/mmm in the tetragonal system. At the same time,
the five elements (cqfl, ~, J, ~) form a basis for the space
of the second-order elastic constants of a hexagonal
crystal.
By continuing the computations as outlined above,
we obtain sixteen linearly independent tensors of rank
four whose expressions are given in Table 3, together
with the symmetry classes for which each is an invariant. Table 4 exhibits the use of these tensors to form
five hierarchies of bases which span the space of second-order elastic constants for each symmetry class,
together with the multiplication tables. The inner product of two tensors A~j~a and B ~ which enters the
multiplication table is defined as (A,B)=A~Ie~B~et.
Bases which span the space of elastic constants of
orders higher than the second may be computed in a
similar manner. The tedious but straightforward algebraic procedures can best be handled by a computer
using available symbolic manipulation programming
such as F O R M A C of 7090/94 IBSYS. In fact, the
third-order elastic constants listed by Hearmon (1953)
have been reproduced by a F O R M A C program but
with the additional feature of exhibiting hierarchies of
bases similar to Table 4. As the number of the thirdorder elastic constants is so many, as evidenced in
Table 1, details of the computer results will not be
given here.
3. O r t h o n o r m a l bases
By a linear combination of elements in a basis in
Table 4, one can, instead, construct an orthonormal
basis whose elements are (1) mutually orthogonal, i.e.
(A,B)=0, if A ~ m # B ~ t , and are (2) normal; i.e.
(A,A) = 1, for elements A and B in the orthonormal
basis. Twenty orthonormal tensors have been constructed. They are displayed in both their tensor forms
and their matrix forms in Table 5, together with the
Table 3. Tensors of rank four and their invariant properties
Expression
Invariance
Isotropic
f
Cubic, tetragonal,
Orthorhombic, monoclinic
Tetragonal, hexagonal,
Trigonal, orthorhombic,
Monoclinic
7C~flcl = (~IgJ2 + gi23Jl) (~/cl ~t2 + ~k2gtl) + (g~2~J3 + gi3~J2) (~k2gl3 + ~/c3612)
+ (&l~j3+ &3~JD (&l&3 + &3&O
= ~0~k3313 + ~i3~J3~/¢l
= (dii2JJ3 + Ji3JJ2) (~/c2~13+ J/c3~t2) + (~il ~J3 + ~i3diJ1) (Jkl ~t3 + ~/c3~tl)
= (&13J2+ &2g~l) (&l&3 + &3& l) + (&l gj3 + &36jl) (&l& 2+ &z&l)
Trigonal
Trigonal (classes 3, 3)
+ (3i2~J3 + ~i3¢5J2) (3/~13t 1 -- 6/c2gt2) + (gil ~Jl -- c~i23J2) (~/c2~t3 + 3e3~t2)
Wight = (&lgJ3 + &36jl) (&l&t - &2&2) + (&13~x - &2g~2) (&l&3 + &3&0
(D~jlc~ =
-- (J*l~2 + g*2J~l) (J/c2~t3 + Je3~t2) -- (J~23J3 + J~3c~2) (J~c~Jt2 + J/c2eStI)
= J~IgJ 1~/c1~/1
Oi~ct = (~23~3 + g/3~2) (6~/¢2~13+ 3/c3~12)
= (3¢3~J33/c13tl + 3,1~113/¢3~t3)
= (~lg~3 + g~33Jl)~/¢3613 + gi3~3(~/cl~t3 + 3/c3~l 1)
TtJIcl
Tetragonal (classes 4,4,4/m)
(&a3J2+ &23~0 (&~&~- &2&2)+ (&~n - &2g~2)(&l&2 + &2& ~)
= (&1~3 + &3g¢~)&~&~+ &lSJl(&l&3 + &3&l)
= (&xg~3+ &3~¢1)&2&2+ &23¢2(&~&3+ &3&~)
}
Orthorhombic, monoclinic
Monoclinic
YIH-O
s y m m e t r y classes for w h i c h e a c h is a n invariant. T a b l e 6
exhibits t h e five h i e r a r c h i e s o f o r t h o n o r m a l t e n s o r
bases w h i c h s p a n t h e space o f t h e s e c o n d - o r d e r elastic
constants.
A n y elastic t e n s o r o f r a n k f o u r possessing certain
c r y s t a l l o g r a p h i c s y m m e t r y m a y be d e c o m p o s e d ac-
TU
277
c o r d i n g to o n e o f t h e s c h e m e s in T a b l e 6. T h u s t h e
elastic stiffness t e n s o r o f crystals o f classes 3,5, for
e x a m p l e , can be r e p r e s e n t e d in the f o r m
VII
CiJkt = •
(c, A K ) A ~ k l ,
(7)
K=I
Table 4. Five hierarchies o f tensor bases and multiplication tables
I
_
Trigonal (3,~)
Trigonal (32, 3m,3m)
Hexagonal
Isotropic --!
(a)
9
6
24
y
1
2
6
4
0
0
0
8
0
0
1
2
8
0
0
0
0
0
0
0
0
16
0
16
8
gt
Tetragonal (4, 2f,4/m)
Tetragonal (422, 4mm,-42m, 4/mmm) - - Cubic . . . . .
I
~- Isotropic --]
]
I
----
(b)
I
p
9
6
0
1
6
0
0
24
12
12
2
0
4
0
8
8
0
0
1
2
8
0
0
8
0
0
0
8
. . . . . . . . . . . . Tetragonal (4,:g, 4/m)
Tetragonal (422, 4mm, :g2m,4/mmm) - - .
-----Hexagonal
- - Isotropic - oc
p
y
5
;g
----
CO
(c)
9
6
24
1
2
6
4
0
8
0
12
0
0
1
2
8
0
0
8
0
0
8
12
0
0
0
0
8
Monoclinic
Orthorhombie
- - - - Tetragonal (422, 4mm, ~2m, 4/mmm) - Hexagonal
Isotropic --]
(d)
II....
9
6
24
0
2
u
z
1
2
6
4
0
8
0
12
1
2
2
0
0
4
0
0
0
0
0
0
0
0
1
2
8
0
0
8
0
0
8
0
0
0
0
2
0
0
0
4
0
0
0
0
0
0
0
0
0
0
0
0
12
0
0
4
0
0
0
0
1
0
0
0
0
0
0
2
0
4
0
0
0
0
0
0
0
0
0
0
0
4
0
4
0
0
4
8
278
THE
DECOMPOSITION
OF AN
ANISOTROPIC
ELASTIC
TENSOR
Table 4 (cont.)
Orthorhombic
I----- Tetragonal (422, 4mm, ~2m, 4/mmm)
Cubic . . . . . .
(e)
Monoclinic
-a
0
2
tt
r
ff
2
0
0
0
2
0
0
2
0
4
4
0
0
4
0
0
4
0
0
0
0
0
0
0
0
0
8
0
0
0
0
0
0
0
0
0
0
4
0
0
0
0
0
0
0
0
0
0
0
4
0
0
0
0
0
0
0
0
0
0
0
0
4
It
9
6
24
0
12
12
1
2
0
6
4
0
0
8
8
1
2
0
1
2
8
0
0
8
0
0
0
1
(111000)
Table 5. Orthonormal tensors of rank four, their matrix representations and their invariant properties
Name
AI
Symmetry
classes
Isotropic
Tensor
A,m I = ½cqj~t
Matrix
1
1
Isotropic
Ac
Cubic
Tetragonal
Orthorhombic
Monoclinic
AHI
Tetragonal
Hexagonal
Trigonal
Orthorhombic
Monoclinic
h H2
Tetragonal
Hexagonal
Trigonal
Orthorhombic
Monoclinic
A H3
Tetragonal
Hexagonal
Trigonal
Orthorhombic
Monoclinic
A TH
Tetragonal
Orthorhombic
Monoclinic
1
A~m II= 61/~ [3flij~t-2~,j~t]
[5rc,~k~
2-~
- 3flim + 2~ij~z]
1
Acmnl= ~
[15ytj~z
- P~m - oc~jkz]
A~jlct H2 = ~2[9&m
- 15y,m +fl,m
- 5~,m]
AIjkt Ha = ¼[2e#la
- & m + 3y,j~t
-fl~m + 0c~m]
AtHct TH
1
A l j I I = 6~5
AijC=
1
V
30
1
A I j HI
°V3
A H H 2 = x_~
AIjH 3 =¼
1
=
0
0
0
0
--24 - - 2 --22
0
0
0
--2 --2
4
0
0
0
11
1 --2
00
00
00
1
0
0
0
0
0
0
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
- 2 1 - - 21
1
Aij~t c =
0
0
0
0
0
l
AII
1
1
AzJI=½
1
1
-2~/2[27rij~t
AIjTH =
- eijla - &jlct + 3yi~t
- f l t m + cti~t]
I
0
0
0
0
0
0
3
0
0
0
0
0
0
3
0
0
0
0
1
0
0
3
-- 13
-- 31 -- 11
00
00
00
- 1 - 1 12 0 0
0
0 0
0 - I
0
0
0
0
0 0 -1
0
0
0
0 0 0 -I
1
1
-3 -5
5 -3
4
4
0
0
0
0
0
0
4
4
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1
0
0 )
0
0
0
0
1
1 --1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
I
0
0
0
0
0
0
I
0
0 --I
( - 1 1 --11
I(_11oooo)
0
0
0
O
0
O
0
1
0
0
O
0
-2-1/2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
O
0 1
0
0
1
YIH-O
TU
279
Table 5 (cont.)
Name
ATCl
Symmetry
classes
Tensor
Matrix
1
Tetragonal
Orthorhombic
Monoclinic
AIjTCl
+ ½ram -
½/hm]
1
A~:~:c2 = 2Vg [3&m
-- 6y~Ikt - n~lkt
ATC2
Tetragonal
Orthorhombic
Monoclinic
ATC3
Tetragonal
Orthorhombic
Monoclinic
AUlct TC3---
ATET
Tetragonal
(classes
4,74,4/m)
At]klTET = -21/2 cotl~z
ATR1
Trigonal
l
1/6
---
AIjTC2
--_
1
A~]kl T R1 ---¼~]Icl
00
2
0
0
0
00
0
0
0
0
00
0
0
0
0
00 1
0
0
0
0
I - 20 - 2 0
11
00
00
00 1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
I
.A_
2~/3
00
00
00
00
00
00 1
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
1
........
00
0
0
0
00
0
0
0
00
0
0
0
00
0
0
0
00 --11 1
0 0
0 0
0 0
1 --1
0
0
0
0
0
0
0
0 1
0 --1
0 0
0
0
0
0 )
0
0
AIjTC3= ~21/--6
AtjTET =
0
0
0
21/2
AzjT m = ¼
I
A Ta2
Trigonal
(classes 3,7)
AIJk.l TR2
= ¼glf]IcZ
0
0
0
0
0
1
1
21/6 [3e~m- 2ntm]
- 10 - 1
0
0
0
0
Azj~,R2 =¼
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1 --1
Orthorhombic
Monoclinic
A~]~t TM = ~
1
[2OIik~
+ ½n,.ua+ 7*.~g~
-½P,m]
A VIII
Orthorhombic
Monoclinic
AIX
Orthorhombic
Monoclinic
A~mIX= 2~2- [20~m-eo~d
Monoclinic
A~jk~x = ~ -1 l~m
A,lgz vIII = ½[2o'~lkt
1
AxaVII = --~
•
AIjIX =
~
AtaX = 2]/2
00
0
0
0
00
0
0
0
00
0
0
0
00 1
0
0
0
0
0
0
0
0
0
1
0
0
0
0
00
0
0
0
0
00
0
0
0
0
00 1
0
0
0
0
I
00
0
00
0
00
0
00
00 1
0
0
0
0
0
I
Ax
01 - 1 0
0
0
0 0
0 0
00 00 - 1
1 -1
0 0
0 0
0 0
1
1
0
0 1
0-1
0
0 0 0
0 0 -1
0
0
0
0
0 --1 0 0
I
AiavnI = ½
- &m + 2?u~z]
1
0)
1 --1
0
0
0
Avn
0 0
1 0
0--2
00
0
0
0
0
0
0
0
00
00
00
00
00
00 1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1 0
0 --1
0 0
0
0
1
0
0
0
0
0
0
1
0
0
280
THE
DECOMPOSITION
OF
AN
ANISOTROPIC
ELASTIC
TENSOR
Table 5 (cont.)
Symmetry
classes
Name
Tensor
Monoclinic
AXI
AXII
Aim x~ = {piye~
Monoclinic
AXlII
Matrix
A z,,-xI = ½
Aim xII = ½~m
Monoclinic
A zJ T M = ½
t
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
(oo
(oo
A idXii I =½
A~flct XIII = ½~i.Ucl
o
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
)
\
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
\
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
\
,
)
T a b l e 6. F i v e h i e r a r c h i e s o f o r t h o n o r m a l b a s e s
AIX
Ax
AXI
AXlI
AXlII
AIX
hx
AXI
AxII
AxIII
AI
All
Ac
ATCl
A~'C2 ATc3
AvII
AVlII AiX
Isotropic --I
Cubic
Tetragonal (422, 4ram, 7~2m, 4/mmm) -- Orthorhombic . . . . . . . . . .
Monoclinic
AX
AXI
AXii
AXm
AI
(a)
--
(b)
--
(c)
--
(d)
AII
AIII
Air
Av
AI
All
An1
AH2
AH3
Isotropic --I
Hexagonal . . . . . . . . . . . .
Trigonal (32, 3m, 3m)
Trigonal (3,~
AVI
AVlI AVlII
A~R1
A~2~2
......
A I All
Ac
ATCl
ATC2
ATC3
Isotropic --I
|
Cubic
12
Tetragonal (42 , 4mm, 7~2m, 4/mmm) - Tetragonal (4, ~[,4/m)
ATET
AI
All
An1
AHZ
An3
ArU
Isotropic --)
|
Hexagonal
Tetragonal (422,4mm,~2m,4/mm) - Tetragonal (4, ~, 4/m) -
ATeT
AI
All
Anl
AH2 An3
ATn
Isotropic --I
Hexagonal
--(
- - Tetragonal (422, 4mm, 2~2m, 4/mmm) - - - Orthorhombic
--
i
i
i
AVII
AVm
I
I
....... Monoclinic
(e)
--
w h e r e AK, K = I , I I , . . . , V I I are the o r t h o n o r m a l basis
a c c o r d i n g to T a b l e 6 (a), a n d (c, A n) is the i n n e r prod u c t o f cty~t a n d A~k ~. Likewise, the elastic stiffness
t e n s o r o f classes 4,4 a n d 4 / m crystals m a y be represented in the s a m e f o r m as (7) b u t w i t h A K c h o s e n to
be t h e o r t h o n o r m a l basis a c c o r d i n g to either (b) or
(c) o f T a b l e 6.
I n o r d e r to facilitate the c o m p u t a t i o n o f i n n e r products o f a n a r b i t r a r y elastic stiffness t e n s o r c~j'~a a n d
a n e l e m e n t o f a n o r t h o n o r m a l basis, T a b l e 7 lists the
Y I H - O TU
Table 7. Formulas f o r the computation o f inner
products in terms o f matrix elements
3
$1 =z~ CII,
1=1
(c, A~)
6
S2=z~ $II,
1=4
Table 7 (cont.)
3
S1 ~-X cii ,
I=1
3 3
S3=X X clj
1=1 J = l
6
S2 ~--,~' sii ,
1=4
3 3
S 3 = X X Clj
I=1 J = l
1
(C, A TC2) = - - ~ [3(c13 + c23) + S1 -- $3]
=~$3
1
(c, A II)
= ~
(c, A c)
-- t/Td [4S2- 3S1+ $3]
[3S1 + 6S2 -- $3]
(c, A H1)
= 61/51 [15c33 -- 2S1 -- 4 S z -- $3]
(C,
(C,
(C,
(¢,
A TC3) = V2"/-3[C44+ C55 __ 2C66]
A TET) = 1/2(c16 -- c26)
A TR1) =2C56+C14--C24
h TR2) __ C15 -- C25 -- 2C46
1
(C, Z TM) = H [Cll--C22]
1
(C, A H2)
(c, A H3)
= ~A-~[18(c13 + C23) + 3c33 + 2S1 + 4S2 -- 5S3]
= ¼18(c44 + c55) -- 2(c13 + c23) + c33 -- 2S1 - 4S2 + $3]
1
(C, A TH) = ~ 2 - [4C66-- 2(C13 + 2C23) + C33 -- 2S1 + S3]
--r
281
- -
1
(C, A ~'cl) = .--7713c33--S1]
go
formulas in terms of matrix elements C l J , I~ J = 1,2,.., 6.
However, for the elastic compliance tensor s~jet, the
coefficients in the formulas must be divided by 2 when
either I or J of s i j is 4, 5, 6 and by 4 when both !
and J of szj are 4, 5 or 6.
I
ClJ :
8.680
0.704
1.191
- 1.804
0
0
0.704
8.680
1.191
1-804
0
0
1.191
1.191
10.575
0
0
0
We now define and denote the norm of ctm by
Ilcll= ( c ~ m .
c~m} ~ •
(8)
When cijgz is in the space spanned by the orthonormal
basis {AK}, it is easy to see that
2: (c, A K ) ) * .
(9)
llcll={ K
The nearest isotropic tensor, denoted by Cakl,° of C~gZ
is therefore
xi
Cljk ---- ,~. (c, A K)A gkt
(10)
K----I
I
cxj =
10.250
2"726
-0"803
0
-0.929
0
2"726
8.680
-0.831
0
0.388
0
-0.803
-0"831
12.992
0
-0'577
0
which has a norm
II
Ilc011=(Kz= I (cO,A~0}*.
(11)
In a similar manner, with respect to the tensor cim,
the nearest tensors of other symmetry classes within
the" class spanned by the basis {AK} may be read off
readily from the representation, and their norms may
be computed according to (9).
(c, A vIII)
(C, A IX)
(C, A x)
( c , A xI)
(c, A xII)
(c, A x u I )
= c13 C23
= [/2[c44-- c55]
=2l/2C46
=2c35
=2c~5
= 2c25
-
-
4. E x a m p l e
According to McSkimin, Andreatch & Thurston
(1965), the elastic stiffness matrix cij at room temperature of quartz (class 32) is given by
-1.804
1.804
0
5.820
0
0
0
0
0
0
5.820
- 1.804
0 \
0
0
0
-1.804
3.988
)
(12)
in 10al dyn.cm -2. Using scheme (a) of Table 6, the
above matrix may be represented in the form
eIj-- 11.368A I+21-386A II+ 0.457A HI
+1"870AHZ+3"663AH3--7.216A Tm .
(13)
For an AT-cut quartz plate ( y x l ) 35 °, if we orient
the Cartesian coordinates in such a way that the (xl, Xz)plane lies on the middle plane of the plate and Xz]]X,
the matrix (12), referred to the new coordinates, becomes
0
0
0
2'895
0
- 0.243
-0.929
0.388
-0.577
0
3.825
0
0t
0
0
-0.243
0
6-912
(14)
which possesses the symmetry of the monoclinic system. This matrix may be represented in either of the
following two ways according to scheme (d) or (e) of
Table 6.
czj = 11-368A I + 21.386A II
f 3"160A m - 3"549A H2- 3"560A H3 + 5"010A TH
1
+ ~
t - 1"301Ac +2.880A TCl -4"091A TCZ -5"800A TC3
+ 1"110A vII + 0-027A v I I I - 1"315AI x - 0"689A x
- 1"155A x I - 1"859A xlI + 0"776A xlII .
(15)
282
THE DECOMPOSITION
OF A N A N I S O T R O P I C E L A S T I C T E N S O R
The nearest tensors of higher symmetry may be read
off from equations (13) or (15). In particular, the
nearest isotropic tensor, C~jkZ,°being an invariant, has
the following matrix:
I
cO=
10.165
0.601
0.601
0
0
0
0'601
10.165
0.601
0
0
0
0.601
0.601
10"165
0
0
0
in a perturbation scheme of plane stress problems in
anisotropic theory of elasticity.
The author is indebted to Dr R.A.Toupin for his
criticism and help.
0
0
0
4.782
0
0
0
0
0
0
4.782
0
0 t
0
0
0
0
4.782
(16)
•
References
Define the number e0 by
e0= Ilcll-IIc°ll
Ilcll
"
(17)
e0 is a scalar constant independent of the rotation of
the axes. It is a measure of 'nearness' of the nearest
isotropic tensor. The value of e0 for quartz at roomtemperature is found to be 0.054. The use of e0 as a
perturbation parameter is currently being investigated
GAZIS,D. C., TADJBAKHSH,I. & TOUPIN,R. A. (1963). Acta
Cryst. 16, 917.
HEARMON,R. F. S. (1953). Acta Cryst. 6, 331.
INSTITUTE OF RADIO ENGINEERS(1949). Proc. lnst. Radio
Engrs. N.Y., 37.
McSKIMIN, H. J., ANDREATCH, P. & THURSTON, R. N.
(1965). J. of Appl. Phys. 36, 1624.
SMZTH, G. F. & RIVLIN, R. S. (1958). Trans. Amer. Math.
Soc. 88, 175.
WEYL, H. (1946). The Classical Groups, Their Invariants and
Representations. Princeton Univ. Press.
Acta Cryst. (1968). A24, 282
The Atomic Mechanism of the Body-Centred Cubic to e-Phase Transformation
BY W. J. KITCHINGMAN
Metallurgy Department, University of Manchester Institute of Science and Technology,
Sackville Street, Manchester 1, England
(Received 12 June 1967)
An atomic mechanism for the body-centred cubic to a-phase transformation is suggested. Atomic
movements over small distances in the [1 lI]b.o.c, direction take place leading to the formation of a new
layer structure. The transformation is completed by rotation of alternate layers of hexagons within
zones related to kagom6 tile structures. The mechanism suggests that certain groups of atoms are more
strongly bonded in the [1 IT] direction than others. The mechanism also suggests that the body-centred
cubic phase exhibits partial long range order prior to the transformation. The ductility of B-uranium
and the brittleness of FeCr and 2NbA1 alloys is discussed in terms of the ordering and coordination
numbers of the atomic positions in the tr structure.
Introduction
The occurrence of the a phase and its properties have
been reviewed by Hall & Algie (1966). The a phase
A
B
always contains at least one transition group element.
In alloys it is one of a series of phases occurring with
the passage from the more open body-centred cubic
structure to the closer packed hexagonal and face-
C
A+B +C
Fig. 1. The a-phase structure described as a layer structure of kagom6 tiles and diamond nets. The outline of the unit cell is
also shown.