Carbon Vol. 18. pp. 147.

153
0 Pergamon Press Ltd.. 1980 Pnnted m Great Bntain

KINETICS OF THE COKE SHRINKAGE PROCESS

DURING CALCINATION
R. W. WALLOUCH and F. V. FAIR
Technical Department. Airco Speer Carbon-Graphite, P.O. Box 828, Niagara Falls. NY 14301. U.S.A.

(Received 30 September 1977)

Abstract-The dimensional changes of raw isotropic and anisotropic coke occurring during calcination
at 50&12Oo”C are examined under isothermal conditions using a high temperature dilatometer. The
isothermal dimensional analysis is adapted to the coke shrinkage process by placing a coke sample into
a dilatometer and measuring the expansion and contraction of the coke as a function of temperature and
time. The apparent activation enerR,y of the coke shrinkage process is computed from the experimental
deformation rate data.

1. INTRODUCTION
Raw cokes undergo chemical and physical changes
during calcination. The correlation between the calci-
nation conditions of temperature and time and
coke properties is well established and has been
reported elsewhere [ 1.21. However, little attention has
been given to the rate of coke shrinkage and total
contraction during calcination. In the manufacture of
carbon artifacts coke porosity adversely affects both
the binder requirements and the apparent density of
the carbon product. One of the principal factors con-
trolling the pore structure and density of the coke is
the interaction between the degree of degasification
and the degree of the coke contraction during calcina-
tion. In consequence, the contraction of raw cokes
was, examined by a dilatometric method under iso-
thermal conditions between 500 and 1200°C as a
function of time. The recorded shrinkage curves were
evaluated by the conventional isothermal analysis and
the Dorn method [3]. The results obtained by the two
evaluation methods are in agreement and confirm the
thermally activated nature of the coke shrinkage pro-
cess during calcination.
2. TEST MATERIAL PREPARATION
Two raw cokes of different origin and isotropy were
used in this study. The starting material for the prep-
aration of the isotropic coke was a coal-tar pitch with
a cube-in-air softening point of 103°C and a Conrad-
son coking value of 56%. Large lumps of raw coke
were obtained by coking 15 pounds of the coal-tar
pitch in an inert atmosphere at 500°C top tempera-
ture. The anisotropic coke was raw delayed petroleum
coke which has been handpicked at the coker.
The large irregular pieces of the isotropic and ani-
sotropic coke respectively were cut into test samples
1 in. in diameter and 5 in. long by means of a core
drill. The change in length of the coke samples with
temperature and time was monitored up to 1200°C by
a sensitive recording dilatometer.
147
3. EXPERIMENTAL
If a raw coke is heated in a dilatometer from
ambient to its top coking temperature of approx.
500°C a reversible thermal expansion is recorded.
Figure I shows the percent length change vs the
reheating temperature of an anisotropic petroleum
coke. The curves u. h and c refer to coke samples
which have been calcined at 700. 800 and 1200°C.
respectively. The curves a and h show that the raw
coke first expands when reheated to the top tempera-
ture attained during the previous heat treatment.
Upon further heating, at a heating rate of 30”C/min.
the reversible expansion changes into an irreversible
contraction.
The irreversible character of the coke shrinkage is
reflected in a lasting contraction at room temperature.
Referring to Fig. 1. it will be noted that the total
expansion decreases for cokes of higher initial calcin-
ing temperature. Upon cooling, the coke samples con-
tract in a manner characteristic for temperature stable
solids. At calcining temperatures above 12OO‘C the
coke shrinks at a rate not readily detectable by the
dilatometer.
In calcining, hydrogen is released along with low
molecular weight aliphatic hydrocarbons. The
amount of hydrogen released during thermal degasifi-
cation of the test cokes has been measured in an inert
atmosphere at various but constant temperatures by
suspending a coke sample (I in. dia. x 1 in.) from an
analytical balance into a vertical tube furnace for
22 hr. Subsequently. the heat treated coke samples
were analyzed for their hydrogen content and helium
density.
The hydrogen analysis was conducted by Clark
Laboratory. Urbana, Illinois using a microanalytical
combustion method. Figure 2 illustrates the hydrogen
content and helium density, respectively. vs calcining
temperature. The percentage change of the hydrogen
content with hold temperature is the same for the two
cokes and follows along a common curve. The helium
148 R. W. WALLOUCHand F. V. FAIR

t 0.6

P’
P’
6- P’
i lZT.
& _,A-’
-08.
_o.9 &_&__&‘d---A---’
,
20 100 200 300 400 500 600 700 600 900 1000
REHEATING TEMPERATTURE, ‘C
Fig. 1. Effect of top coking temperatures a = 7OOT, h = 8OO”C,and c = 1200°C upon the dimensional
changes of anisotropic coke in reheating.

density of the heat treated petroleum coke and of the
pitch coke increases nearly linear with temperature.
The thermal removal of carbon in the form of vola-
tile hydrocarbons causes an average weight loss of
4.5%. Figure 3 illustrates the percent weight loss and
change in volume, respectively, vs calcining tempera-
ture. It will be noted that both cokes have practically
the same weight loss when heated to 900°C in incre-
ments of 100°C and after a retention time of 22 hr at
each temperature. The volume contraction of 18% for

3.5 -
0 - 1.9

- 1.6

- 1.5

- 1.4
A PETROLEUM COKE
0.5- 0 PITCH COKE

.
500 600 700 600 900
CALCINING TEIwIPERATURE,~C

Fig. 2. Percentage change in hydrogen content and helium density of petroleum and pitch coke between
500 and 900°C.
 Kinetics of the coke shrinkage process 149

CALCINING TEMPERATURE ,“C

500 600 700 800 900

-2 -
-4 -

-6 -

-a -

-10 -
-12 -

-14 -
J
-16
9
-18

-22 a PETROLEUM COKE

-24 c o PITCH COKE I

Fig. 3. Percent weight loss and volume contraction of petroleum and pitch coke between 500 and 900°C.

h L . I . . . . * 1 1 . I

I 2 3 4 5 6 7 6 9 IO II I2 I3 I4

CALCINIW TEWERAlU#. lC X IO-*

Fig. 4. Dimensional changes of anisotropic coke (Curve I) and isotropic coke (Curve 2) during non-
isothermal heating. Heating rate: 1507C/hr~1200’C.
150 R. W. WALLOUCHand F. V. FAIR

TIME (MINUTES)

c
30 60 90 120 I!50
I

676C
-I

-2

-3

c
4 -4

%
3f --c

i -6
J

-7

-8

-9

-10
Fig. 5. Plot of percent isothermal shrinkage vs time at constant temperatures of isotropic coke.

anisotropic petroleum and 22% for isotropic pitch
coke, respectively, observed at 900°C is the principal
factor controlling the coke density.
The magnitude of dimensional changes observed on
raw coke samples (1 in. dia. x 5 in.) during non-iso-
thermal heating is illustrated in Fig. 4. Curve 1 refers
to the anisotropic coke and Curve 2 to the isotropic
coke. These curves show that raw cokes expand when
heated up to 500-600°C and then, upon further heat-
ing, at a rate of 150°C per hour, the expansion
changes into a contraction. A precise evaluation of
the non-isothermal shrinkage curves illustrated in
Fig. 4 is difficult since the time at which shrinkage
begins is not well defined. However, if the shrinkage
process of cokes is studied under isothermal condi-
tions, a quantitative analysis of the experimental iso-
therms is possible. The isothermal dimensional analy-
sis technique was used first by Dorn[3] to study the
kinetics of creep of metals and later was adapted by
Bacmann and Cizeron[4] to the sintering process of
uranium dioxide compacts.
Following the Dorn method, a coke sample was
permitted to contract in the dilatometer at a constant
temperature T,. Subsequently, the temperature was
raised rapidly to a higher temperature T, and the
sample was allowed to contract further at an acceler-
ated rate. Figure 5 shows the effect of rapid tempera-
ture changes of approx. 70°C in 30-min time intervals
on the length of a sample of isotropic pitch coke.
Temperature equilibrium was attained after 3-4 min
elapsed time. The inflections which appear on the
curve characterize points where the isotherms under-
went a change in rate due to the rapid increase in
temperature. The rates of contraction of two adjacent
isotherms at the point of temperature change are
defined by the slope of the curves.
The method enables one to determine the apparent
activation energy (temperature factor) with a single
sample and therefore eliminates parameters which
might vary from one sample to another. However, no
information is obtained by the Dorn method about
the pre-exponential factor in the Arrhenius equation
which reflects differences in the shrinkage rate of
cokes of different anistropy.
4. THEORY
The isothermal shrinkage curves of the type shown
in Fig. 5 are adequately described by eqn (I), which
has been found to be representative for many solid-
state kinetic pracesses[4].
P = A. t”. (1)
 Kinetics of the coke shrinkage process 151

The proportionality factor A and the time exponent n 5. RESULTS AND DISXSSION
are temperature-dependent material constants. The For the computation of the apparent activation
shrinkage parameter P represents the fraction of non- energy AE it is required to know the average time
contracted coke at any given time t and temperature. exponent narg and the factors A, and A2 correspond-
From Fig. 4, it is evident that a marked coke contrac- ing to T, and T2. The values of II and A are obtained
tion starts in the range 50&6OOC, reaches a maxi- from a plot of log P vs log t. Figure 6 is such a
mum rate of shrinkage in the range 700-800°C and straight-line representation of the five isotherms
attains a nearly zero rate of contraction at 1200°C. shown in Fig. 5. The second point of each isotherm in
Thus. the shrinkage parameter P can be defined by Fig. 5 corresponds to zero time in Fig. 6. The numeri-
eqn. (2): cal values of II and A as a function of temperature arc
defined by the slope and intercept of the linear plots
Al
p=lep in Fig. 6. Table 1 is a summary of the A and )I va.lues
lo - 1, as computed by the least squares method. Referring
where I, is the initial length at room temperature; to Table 1. the average n-value is approx. -0.: for
1I the final length at the end of the heating cycle at isotropic coke and -0.4 for anisotropic coke, indicat-
1200°C; Al = 1, - 1,. where 1, is the length of the ing that the total contraction as well as the rate of
sample at time t. shrinkage is different for cokes of different isotropy.
The rate R of the isothermal shrinkage is defined Table 2 summarizes the apparent activation
by the first derivative of eq. (1): energies as computed by the Dorn method for the two
cokes. It can be seen that the average value of the
time exponent II changes with the isotropy of the
Rate(R) = dPldt = nAt”/t (3)
cokes, whereas the apparent activation energy AE for
From (1) and (3). the two cokes has a value of 76 kcalmol on the aver-
age. The apparent single value of BE for the two
Rate(R) = nA”“.P1-“” (4) cokes suggests that the rate-controlling mechanism of
or the coke shrinkage process is the same for isotropic
and anisotropic coke.
In Rate (R) = In n + (l/n) In A + (1 - l/n) In P. (5) Figure 7 shows a plot of log Rate vs 17 for the
two cokes, where log Rate is given by [ 1 II log .d +
We assume that the “structural configuration” of the log n]. The slope of this plot equals Q/nR and yields
coke remains unchanged at the point of rapid tem- the apparent activation energy AE = Q,n. More im-
perature change. Mathematically, this means that, for portantly, it is seen that the lines corresponding to the
two adjacent isotherms, P, = P, and as an approxi- two cokes are nearly parallel to each other indicating
mation n, = nz. that the activation energy is apparently the same for
This being the case. the temperature-dependent both isotropic and anisotropic cokes and is found to
terms in eqn. (5) are separated from the temperature- be about 71 kcal;mol. However, in using the empirical
independent terms by forming the ratio of the shrink- eqn. (1) for describing the kinetics of the coke shrink-
age rates R, and Rz of two adjacent isotherms and age process, certain assumptions had to be made
one obtains eqn. (6): regarding the value of n and A over the temperature
range from 500 to 1200°C. If one of the material par-
ln(RateR,/RateR,) = (l/n)ln(A,/A,). (6) ameters varied with temperature, another mechanism
could become rate controlling. Such a possible super-
It is assumed, see Martens et al. [S] that the factor A
position of two competing processes necessitates one
has an exponential temperature dependence similar
to accept with reservation a single value of AE for the
to that valid for rate and is given by eqn. (7):
two cokes. The agreement of the results obtained with
A = A,e-QIRT. (7) the Dorn method and the Arrhenius plot, confirms
the validity of eqn. (1) for describing the kinetics of
Substituting (7) in (6) one derives eqn. (8):
isothermal shrinkage of raw coke during calcination.

(Iln)ln(A,;A,) = -g(l/T, - l/T,) (8) 6. SUMMARY AND CONCLLSIOUS

The irreversible contraction during calcination of

where according to Beck[6] Q,‘n is the apparent
raw cokes of different origin and isotropy was exam-
activation energy AE of the isothermal shrinkage
ined in a dilatometer under isothermal and non-
process, or from eqn. (8),
isothermal temperature conditions. The parameters
obtained for two different cokes by two different
AE = Q/n = 2.303. R ~. I/navg1% (AdA 1)
Tz - T, methods are compared. The interaction between the
rate of degasification and coke contraction during cal-
(9) cination are believed to be the principal factors con-
where II,_ = (n, + n,jj2. trolling the pore structure and density of cokes. The
152 R. W. WALLOLJCH and F. V. FAIR

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5
LOG t (MINUTES)

Fig. 6. Plot of log P vs log t for various isotherms of isotropic coke.

Table 1. Factors A and n for isotropic and anisotropic coke

Isotropic Coke Anisotropic Coke

Calcining Structural Time No. of Structural Time No. of
Temperature Factor, Exponent, Data Factor. Exponent. Data
l i°CI A n Points A n Points

676 13. 1647 -0.1475 19 15.7475 -0.2762 18

746 8.0077 -0.1706 17 6. 1561 -0.3974 14

016 5.4364 -0. 1764 12 2.5317 -0.3136 6

584 3.6158 -0.2092 10 1. 1442 - 0. 5732 4

952 2.2819 -0.2887 7

Average Time

Exponent, nava -0.1985 -0.3901

Table 2. Activation energies computed by the Dorn method eqn (9)

isotropic Coke Anisotropic Coke
Isothermal Average Time Structural Activation Average Time Structural Activation
Temperatures Exponent. Factor, Energies,AEa Exponent, Parameter Encrgios.AEa
(OK) AT nave log -Q./Al (kcallmol) na,g. ~opl AZ/AI (kcal/nolJ

949 70 -0.1591 0.2159 86 -0.3368 0.4079 77

1019

1019 70 -0.1735 0.1682 70 -0.3555 0.3859 79

1089

1089 68 -0.1928 0. 1771 78 -0.4434 0.3449 66

1157

1157 68 -0.2489 0.1999 77

1225

Average Activation

Enerzies,AEa ?8+7 kcaI/mol 74*7 kcal/mol

 Kinetics of the coke shrinkage process 153

-6.C

-5.C

-4.c

-3.1

-29

- I.(

B
CURVE 2

0.1
\
9

I.{
1.1 LO 0.9 0.e a7
t/T X 103;K-’
Fig, 7. ~rrhenius type plot for isotropic coke (Curve It and anisotropic coke (Curve 2)

shrinkage process upon heating is adequately de-
scribed for the cokes studied by the empirical equa-
tion P = At” with the maximum rate of shrinkage
occurring between 700 and XOOC. Above 5OO.C. raw
cokes shrink upon heating.
The apparent activation energy of the coke shrink-
age process is nearly the same for the two cokes stud-
ied independent of their origin and isotropy and has
an average value of 76 kcal/mol, the relative error
being approx. + 10%. The single value of the tempera-
ture factor seems to indicate that the rate controlling
mechanism of the coke contraction is the same for the
two coke types.
The pre-exponential factor in the rate equation
(eqn. 7) of the shrinkage process varies with the type
and the structural configuration of the initial coke
material. The anisotropic coke had a higher shrinkage
rate than the isotropic coke which leads to higher
values of A2:‘A, and YI.
REFERENCES
I. P. Rhedey. Structural changes of petroleum coke dur-
ing calcination, Trtors. Mc~ttill. Snc. AI,VE 239. 1084
(19671.
2. M. P. Whittacker. F. C. Miller and H. C. Fritz. Great
Lakes Co.. Structural changes accompanying coke
calcination. In&. E~run C/tern. Prod. Rrs. Dw. 9(2:1, 187
( 1970).
3. J. E. Dorn. Crwp LIII(~Recctcery (Edited by R. Maddin).
ASM Publication. Cleveland, Ohio (1957).
4. J. J. Bacmann and G. Cizeron. Dorn method in the
study of initial phase of uranium dioxide sintering. J.
urn. Cer. Sot. S1(4), 209 (1968).
H. E. Martens, L. D. JafTe and D. D. Button. Jet Pro-
5.
pulsion Laboratory Progress Rep. No. 20 373 (19%).
6. P. ,A. Beck. .A&. Plr_rs.3. 245 (1954).