B-4CO.V-DEGREE O F ORIENT.4 TION OF GR.

4PHITE 477

A METHOD FOR DETERMINING THE DEGREE

OF ORIENTATION OF GRAPHITE
By G. E. BACON

The preferential orientation of the carbon planes in blocks of graphite prepared by extrusion or
pressing is described in terms of a n orientation function and i t is shown how this varies with the
particle size of the starting materials. An ' anisotropy factor ' is defined as a useful index of the
material. A discussion is given of the effect of orientation on the anisotropy of thermal a n d
irradiation expansion of bulk graphite.

Introduction
Changes in the dimensions of blocks of graphite under irradiation are of major importance
in the design and operation of a graphite-moderated nuclear reactor. When graphite blocks are
produced by extrusion or pressing the resulting product is found to be oriented, that is the
layer planes of the graphite crystallites are distributed preferentially, tending to lie, respectively.
parallel to the extrusion direction or perpendicular to the pressing direction. It is found1 tha%
the dimensional changes of graphite crystallites under irradiation are markedly anisotropic just
as are the coefficients of thermal expansion. As a result the dimensional changes of macro-
scopic graphite blocks will be anisotropic and will be influenced by the degree of preferential
orientation. On account of the porous structure of graphite the fractional increase of macro-
scopic volume on irradiation is much smaller than that of the crystallites, just as it is found2
+
that the mean coefficient of thermal expansion is much smaller than g(uc 2 4 , where a,, a,
are the coefficients of linear (thermal) expansion parallel to the c, a , axes of a crystallite. More-
over, the bulk anisotropy can be accounted for only qualitatively in terms of a,, a, and the
crystallite Orientation. Nevertheless, the measurement of the degree of orientation of macro-
scopic samples is of considerable importance to atomic pile design, particularly if different
methods for graphite preparation are under consideration. The present paper describes t h e
quantitative measurements of orientation by X-ray diffraction methods and gives some data on4
the effect of the nature of the starting materials on the orientation of the product prepared by
extrusion or pressing. The effect of the materials is noticeable even in the production of
cylindrical specimens for X-ray powder photography, by extrusion through glass capillary
tubes, where it is noticed that highly crystallinenatural graphitesgive muchmoreorientation than
relatively poorly crystalline electric furnace graphites.
The results are expressed in terms of an ' orientation function ' which measures the ' density
of layer-plane normals per unit solid angle ' as a function of the inclination to the extrusion axis
or pressing direction.
The paper is based on a previously-classified report written in 1949.

Experimental methods
Specimens of the various graphite saniplcs were prepared in the form of thin plates, about 1
mm. in thickness and 2 cm. square. In all cases, X-ray transmission photographs were taken
using unfiltered Cu K radiation, but, in order to secure optimum conditions, the precise arrange-
ments were somewhat different for the extruded and pressed samples.
For the extruded graphites, the arrangement shown in Fig. l(a) was used, the slice being cut
with its plane parallel to the extrusion direction and being mounted perpendicular to the incident
X-ray beam, with the extrusion direction horizontal. The photographic film was placed t o
record the 0002 diffraction ring from the graphite unit cell. This halo occurs at a Bragg angle
8 of about 13" and its radius therefore subtends an angle of 26" at the specimen. Reflection t o
the top and bottom of the ring takes place only from 0002 planes which are parallel to the
extrusion direction: reflection to the two sides occurs for planes inclined at 90"- 8, i.e., 77" t o
the extrusion direction. Reflection to any angular position y on the photograph takes place for
planes whose normals are inclined at an angle 4 to the extrusion direction, where 4 is given by
cos 4 = 0.974 cos y
The value of 4 given by this expression is plotted against y in Fig. 2(a). It will be seen that the
diffraction ring covers all values of 4 from 13" to 90".

J. appl. Chem., 6, November, 1956

478 BACON-DEGREE O F ONIENTATI0,V O F G R A P H I T E

By means of a microphotometer the intensity of

blackening is measured along radii of the diffraction
ring a t various values of y. The integrated intensity of
blackening across the ring at any point is proportional
to the orientation function I(#),which is defined as the
density of 0002 plane normals per unit solid angle about
the angle #, and hence, using Fig. 2(a), a curve can be
plotted to show I($), as a function of #. Such a curve for
a well-oriented graphite is shown in Fig. 3(a). I t will be
seen that the fact that the diffraction ring does not
cover # values below 13" is of no importance, since with
extruded samples there are few planes whose normals
have such low values of #. Each hexagon of carbon
atoms throughout the sample is regarded to be
FIG.1 . Expevimental arrangements represented by a line drawn through an origin 0 and
for measurement of orientation of
( a ) extruded, and ( 6 ) pressed perpendicular to the plane of the hexagon. I(#)Sw
samples. The aryows, 1:. P will then be the number of these normal-lines
represent the directions of extrzr- which lie within a small solid angle 6 w about the
sioii and pressing resjJrciii~ely
direction #.
In the case of the presstd graphites the arrangement shown in Fig. I(b) is used. The slice
is cut with its plane parallel to the pressing direction and is mounted with the latter in the vertical
plane containing the incident X-ray beam, but tilted from the vertical by 13", away from the
oncoming beam. By this means, the planes which contribute to the bottom of the diffraction
ring are those for which # = 0, i.e., those whose normals are parallel to the pressing direction.
For the pressed graphites # = 0 will, of course, give the maximum value of I(#),since the layer
planes tend to lie perpendicular to the pressing direction.
The variation of # with y , the angle around the diffraction ring, can be shown to be
cos # = cos2 77"-sin2 77" x sin y
= 0.05-0.95 sin y

This relation is illustrated graphically in Fig. 2(b), from which it follows that the diffraction
ring will be asymmetrical about the horizontal, since, whereas the bottom of the ring corresponds
to # = O", the top of the ring corresponds to 4 = 26".
Microphotometry of the films along radii a t various values of y gives the corresponding
values of I(#),by integration across the ring. Using Fig.2 (b),I(#)can then be plotted in terms of
#. Fig. 3(b) shows a curve obtained in this way with a well-oriented pressed graphite.
I n the cases of both the extruded and pressed samples a collimated beam of X-rays emerging
from a pin-hole about 0 . 5 mm. in diameter is used. The specimen is oscillated from side to
side, over its width of 2 cm. during the exposure in order that the resulting photograph may
represent the orientation of an appreciable volume of material, rather than that of only a few
grains.

30 60 90
(a AZtMUTH ANGLE, 1' (41 t
v

FIG. 2. The relation between 6,the angle between the direction of extrusion FIG.3. Curves of I(#) versus
or pressing and the normal to the reflectingplanes contributing aparticular 4 for typical samples of
portion ofthe daffraetion ring*and y the azimuth angle on the photograph well-oriented graphite. Curve
f o r ( a ) extruded. ( b ) pressed samples. T h e experimental arrangement i s ( a ) i s for a n extruded. curve
assumed to be that shown in Fig. 1 (b) f o r a pressed specimen

J. appl. Chem., 6, November, 1956

 U,-lCOA-DEGKEE O F OKIE,VT.-l T I O X O F G R A P H I T E 479

-4riisotropy factor
I n order to have a simple numerical magnitude as an expression of the degree of orientation,
instead of a curve relating I(+)and 4, the concept of ' anisotropy factor ' has been found useful,
particularly when the effect of orientation on the anisotropy of the physical properties of bulk
graphite is being considered.
If therewere no voids in bulk graphite (which is not the case, for the density of bulk graphite
is appreciably less than the X-ray density of the crystallites), then the value in any direction of a
physical property, such as thermal expansion or expansion under neutron bombardment, could
be calculated from the orientation function and the known coefficients of expansion (thermal or
neutron), both parallel, u ~and, perpendicular, oc, to the layer planes, of the crystallites. For a
block of extruded graphite, calculation could then be made from I(+)(assuming no change in
porosity as a result of expansion), of the ratio of bulk expansions in directions perpendicular and
parallel to the extrusion direction. This ratio is called the ' anisotropy factor for expansion' due to
orientation. For pressed graphite the ratio of the calculated expansions along and perpendicular
to the pressing direction is taken. In the case of graphite, a, is negligible in comparison with
U, for both thermal expansion and neutron expansion, with the result that the ' orientation

anisotropy factor ' will be the same for both phenomena. In practice the expansions will
depend not only on the orientation of the crystallites, which we have determined, but also, in a
porous material, on the variation of line density with direction which may be regarded as an
' orientation of holes '. \?;ith extruded graphites this may well cause increased anisotropy of
expansion. The shape and distribution of the cavities between the crystallites will determine
the extent to which the expansions of the crystallites are translated into expansion of the bulk
material.

i
I'IG. 4 . ~lUg?'Ulil f 0 ill'L/SlY& /lLE
calculation Olf the ' aizisofiwfij,
farlov '. The Z axis i s takrii
11s the diiectinii of e s f i w z o i i or
pressiiig
X

The calculation of the anisotropy factor can be understood

from Fig. 3 . Consider the element of solid angle included
hetween the angles 4, #
IJ +
d+ and p, /3 dfi as shown in the +
figure. The magnitude of this element of angle will be equal to sin +.d+.dp. ConsequentlJ-,
the number of graphite crystallites which have their c axes within this angle will be proportional
to
I(+).sin +.d$.dp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...................... . (1)
For a block of material having principal coefficients of expansio a22,a,,, specified for either
thermal or neutron expansion, it can be shown that the coefficient ofjexpansion in a directioii
for which the direction cosines are c,,, c2,, c,:~,relative to the principal axes, is given by
g = a,, c , , ~i- a2*~ 2 A3 a33 ~ c332 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
+
Hence for a graphite crystallite oriented at it follows that the coefficient o
direction 02 is
go, = U, cos2 ++
u, sin2$ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .(3)
and the coefficient of expansion in the direction OX is
+
go, = uccos2a u,sin2a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ' (4
(where a is the angle indicated in F-g. 4.)
For a block of graphite it can be shown from equations ( l ) ,(3)fby;integration with respect
+
to (after expressing a in terms of 4, p), that the coefficient of expansion in the direction 02 is
go, = uc/I(4). cos2 4. sin 4.d4 + 0, fI(4). -.......................
sin3 4.d4 .(5)
./I($). sin $.d4
J. appl. Chem., 6, November, 1956
480 BACON-DEGREE O F ORIENTATION OF G R d P H I T E

and for the direction OX

ac--a, /I(#). sin3 +. d+ ...................................
a,, = a, -+-2
x
/I(#). sin #. d#
If a, is neglected in comparison with a,, as experiment shows can be permitted for
graphite, we find that the orientation anisotropy factor, will be
00,
- = & !I(#). sin s#. d+ ............................................ (7)
0 02 /I(#). cos2 #, sin #.d$
[The limits of integration being ,; 0 in all the above integrals.]
By graphical integration, this ratio can be determined from the I(#) curve for any graphite
sample.
If there is a random distribution of crystallites, then I(#)has the same value for all values
of the angle # and equations (5) and (6) reduce to
aoz= aox= a, + ++ a,, as is well-known.
As an example, it can be shown that if I(#) E sin #, which would approximate to the case of
a slightly oriented extruded graphite, then the anisotropy factor is equal to 1 -5. Similarly, for
more highly-oriented graphites represented in turn by I(#) = sin 2#, sin 4#, sin 6#, sin we cal-
culate anisotropy factors of 2, 3, 4 and 5, respectively. Thus, in order to obtain a large aniso-
tropy of expansion the I(#) curve must be very sharp, corresponding to a high degree of orienta-
tion.

Diffraction photographs of the type described were taken with a number of different
extruded and pressed graphites, made from a variety of starting materials. Details of these,
together with the values of the anisotropy factor obtained from the I(#) curve and graphical
integration of the numerator and denominator of the term in equation (7), are given in Table I.

Table I
-4 )zisotYopY factors of various graphites
KO. Process Starting materials .\nistropy
factor

1 Extruded Ceylon graphite 5

2 70qb Ceylon graphite, 300; coke 2
coke and pitch 1.25
soot and pitch I .01
Pressed coarse Madagascar flake 7
fine Madagascar flake 5
Ceylon graphite 3.5
7006 Ceylon graphite, 300,; coke 3
I

coke and pitch 1

I t can be shown by ordinary X-ray powder photographs of these materials that there is a
continual reduction of the average crystallite size in passing through the series of extruded samples
from (1) to (4). Moreover, the materials can clearly be seen to become less and less crystalline as
judged by the naked eyc or the microscope, varying from the large flakes of natural graphite to
the very fine soot particles. It can be seen from the X-ray powder photographs and the measured
values of the anisotropy factor that the orientation of the extruded block falls off rapidly as the
particle size of the initial materials is reduced. For the natural graphite sample, the orientation
is very marked on the photograph ; for the graphite-coke mixture it is very much reduced but
still clear to the eye, whereas with the pile graphite it is barely discernible except by photometry.
With the soot-pitch mixture the orientation is found to be negligible by photometry.
For the series of pressed samples, Nos. 5-9, there is also a decrease of the degree of orientation
with reduction of crystallite size, varying from the very high orientation of the coarsest natural
griphite to the negligible orientation of the pressed pile graphite.
There will be two determining factors in controlling the final degree of orientation : first it
will depend on the anisotropy of the shape factor, that is, whether the initial material consists
of large flat plates or, at the other extreme, roughly spherical particles. Secondly, it will depend

J. appl..Chem., 8, November, 1956

 B.4COS-DE(;REE O F ORIENT.4TION O F GK=IPHITE 481
on the perfection of orientation within a particle. In practice the two factors are closely related:
thus a natural graphite flake is not only very lamellar in shape but, also, it is built up of graphite
layers stacked with a high degree of parallelism.
I t is of interest t o compare the coefficients of expansion, calculated from I($) in the above
manner for ‘ void-free ’ graphite, for the cases of extruded and pressed samples which give a
particular value of the orientation anisotropy factor.
In the case of extruded graphites uoI is greater than uozand for the various values of the
ratio “0. listed in the first column of Table 11, can be obtained the values of uoz and uoX,as
u02
fractions of uC,applicable to thermal or neutron expansion as appropriate, given in the second
and third columns of the table. These are such that the sum of the smaller value, uo2,plus twice
the larger one, uOx,equals unity.
On the other hand, for pressed graphites, uoz is greater than uox and if the same series of
002
values is taken for the inverse ratio -values for uo2and uoxare obtained as shown in columns
U0.l
4 and 5. These are now such that the sum of the larger value plus twice the smaller one equals
unity. In particular, for a given anisotropy ratio, the minimum expansion in any direction is
smaller for extrusion than for pressing. This arises, of course, because the extruded graphite
has its minimum expansion in only a single direction 0 2 , whereas the pressed blocks have their
(larger) minimum expansion for all directions in the XY plane.

Table I1
Coe//iicieuts of expaiasiotz .for samples of different oriciztatiott anisotropy factor
Orientation Extruded Pressed
anisotropy ___ ____ -__- .__
factor uoz (Along uoX (perpendi- uoZ (along uox (perpendi-
extrusion cular to pressing cular to
direction) extrusion direction pressing
direction) direction)

1..5:1 0.2.5 0.37 0.43 0.28

2:1 0.20 0.40 0.50 0.25
3:1 0.14 0.42 0.60 0.20
5:I 0.09 0.45 0.70 0.14
10: 1 0.05 0.48 0.84 0.08
tc 0.0 0.50 1.0 0.0

It is emphasized that what has been called the orientation function I($), relates solely to the
angular distribution of carbon hexagons in space. I t gives no information about the inter-
crystallite cavities and the extent to which these may limit, in an anisotropic fashion, what can
be called the ‘ free expansion ’ of the crystallites. While this is of course a limitation of the
concept, it is at the same time the strength of this X-ray measurement, compared with measure-
ments such as thermal expansion and conductivity, where the effects of the crystallites and their
interstices are inextricably intermingled. Besides providing a useful index characterizing a
sample of graphite, the ‘ anisotropy factor ’ has been found useful for specifying quantitatively
the way in which, for example, the orientation of an extruded bar varies as one travels outwards
from the axis of the bar to the surface which was in contact with the extrusion die. These same
ideas may, of course, be found useful for other materials than graphite in which there are lamellar
characteristics and anisotropy of physical properties.

Atomic Energy Research Establishment

Hanvell, Berks.
Received 15 March. 1956

References
Bacon, G. E., & Warren, B. E., A d a Cryst., 19.56,i n Nelson, J. B., & Riley, D. P.. Proc. p k y s . S O C .I , o t i k ,
the press 1945, 57, 477

J. appl. Chem., 6, November, 1956